## Problem
Ref: https://github.com/neondatabase/cloud/issues/23461
and follow-up after: https://github.com/neondatabase/neon/pull/10553
we used `echo` to set-up `.wgetrc` and `.curlrc`, and there we used `\n`
to make these multiline configs with one echo command.
The problem is that Debian `/bin/sh`'s built-in echo command behaves
differently from the `/bin/echo` executable and from the `echo` built-in
in `bash`. Namely, it does not support the`-e` option, and while it does
treat `\n` as a newline, passing `-e` here will add that `-e` to the
output.
At the same time, when we use different base images, for example
`alpine/curl`, their `/bin/sh` supports and requires `-e` for treating
escape sequences like `\n`.
But having different `echo` and remembering difference in their
behaviour isn't best experience for the developer and makes bad
experience maintaining Dockerfiles.
Work-arounds:
- Explicitly use `/bin/bash` (like in this PR)
- Use `/bin/echo` instead of the shell's built-in echo function
- Use printf "foo\n" instead of echo -e "foo\n"
## Summary of changes
1. To fix that, we process with the option setting `/bin/bash` as a
SHELL for the debian-baysed layers
2. With no changes for `alpine/curl` based layers.
3. And one more change here: in `extensions` layer split to the 2 steps:
installing dependencies from `CPAN` and installing `lcov` from github,
so upgrading `lcov` could reuse previous layer with installed cpan
modules.
## Problem
We wish to make heatmap generation additive in
https://github.com/neondatabase/neon/pull/10597.
However, if the pageserver restarts and has a heatmap on disk from when
it was a secondary long ago,
we can end up keeping extra layers on the secondary's disk.
## Summary of changes
Persist the heatmap after a successful upload.
## Problem
Hopefully this can resolve
https://github.com/neondatabase/neon/issues/10517. The reason why the
test is flaky is that after restart the compute node might write some
data so that the pageserver flush some layers, and in the end, causing
L0 compaction to run, and we cannot get the test scenario as we want.
## Summary of changes
Ensure all L0 layers are compacted before starting the test.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
In #9895, we fixed some issues where `ClearVmBits` were broadcast to all
shards, even those not owning the VM relation. As part of that, we found
some ancient code from #1417, which discarded spurious incorrect
`ClearVmBits` records for pages outside of the VM relation. We added
observability in #9911 to see how often this actually happens in the
wild.
After two months, we have not seen this happen once in production or
staging. However, out of caution, we don't want a hard error and break
WAL ingestion.
Resolves#10067.
## Summary of changes
Log a critical error when ingesting `ClearVmBits` for unknown VM
relations or pages.
## Problem
We want to switch proxy and ideally all Rust services to structured JSON
logging to support better filtering and cross-referencing with tracing.
## Summary of changes
* Introduce a custom tracing-subscriber to write the JSON. In a first
attempt a customized tracing::fmt::FmtSubscriber was used, but it's very
inefficient and can still generate invalid JSON. It's also doesn't allow
us to add important fields to the root object.
* Make this opt in: the `LOGFMT` env var can be set to `"json"` to
enable to new logger at startup.
This PR does a bunch of things:
* only allow errors of the server cert verification, not of the TLS
handshake. The TLS handshake doesn't cause any errors for us so we can
just always require it to be valid. This simplifies the code a little.
* As the solution is more permanent than originally anticipated, I think
it makes sense to move the `AcceptAll` verifier outside.
* log the connstr information. this helps with figuring out which domain
names are configured in the connstr, etc. I think it is generally useful
to print it. make extra sure that the password is not leaked.
Follow-up of #10640
Refactor how extensions are built in compute Dockerfile
1. Rename some of the extension layers, so that names correspond more
precisely to the upstream repository name and the source directory
name. For example, instead of "pg-jsonschema-pg-build", spell it
"pg_jsonschema-build". Some of the layer names had the extra "pg-"
part, and some didn't; harmonize on not having it. And use an
underscore if the upstream project name uses an underscore.
2. Each extension now consists of two dockerfile targets:
[extension]-src and [extension]-build. By convention, the -src
target downloads the sources and applies any neon-specific patches
if necessary. The source tarball is downloaded and extracted under
/ext-src. For example, the 'pgvector' extension creates the
following files and directory:
/ext-src/pgvector.tar.gz # original tarball
/ext-src/pgvector.patch # neon-specific patch, copied from patches/ dir
/ext-src/pgvector-src/ # extracted tarball, with patch applied
This separation avoids re-downloading the sources every time the
extension is recompiled. The 'extension-tests' target also uses the
[extension]-src layers, by copying the /ext-src/ dirs from all
the extensions together into one image.
This refactoring came about when I was experimenting with different
ways of splitting up the Dockerfile so that each extension would be in
a separate file. That's not part of this PR yet, but this is a good
step in modularizing the extensions.
## Problem
Image layer generation could block L0 compactions for a long time.
## Summary of changes
* Refactored the return value of `create_image_layers_for_*` functions
to make it self-explainable.
* Preempt image layer generation in `Try` mode if L0 piles up.
Note that we might potentially run into a state that only the beginning
part of the keyspace gets image coverage. In that case, we either need
to implement something to prioritize some keyspaces with image coverage,
or tune the image_creation_threshold to ensure that the frequency of
image creation could keep up with L0 compaction.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Erik Grinaker <erik@neon.tech>
## Problem
We don't currently have good alerts for critical errors, e.g. data
loss/corruption.
Touches #10094.
## Summary of changes
Add a `critical!` macro and corresponding
`libmetrics_tracing_event_count{level="critical"}` metric. This will:
* Emit an `ERROR` log message with prefix `"CRITICAL:"` and a backtrace.
* Increment `libmetrics_tracing_event_count{level="critical"}`, and
indirectly `level="error"`.
* Trigger a pageable alert (via the metric above).
* In debug builds, panic the process.
I'll add uses of the macro separately.
## Problem
I noticed when onboarding lots of tenants that the AZ scheduling
violation stat was climbing, before falling later as optimisations
happened. This was happening because we first add the tenant with
PlacementPolicy::Secondary, and then later go to
PlacementPolicy::Attached, and the scheduler's behavior led to a bad AZ
choice:
1. Create a secondary location in the non-preferred AZ
2. Upgrade to Attached where we promote that non-preferred-AZ location
to attached and then create another secondary
3. Optimiser later realises we're in the wrong AZ and moves us
## Summary of changes
- Extend some logging to give more information about AZs
- When scheduling secondary location in PlacementPolicy::Secondary,
select it as if we were attached: in this mode, our business goal is to
have a warm pageserver location that we can make available as attached
quickly if needed, therefore we want it to be in the preferred AZ.
- Make optimize_secondary logic the same, so that it will consider a
secondary location in the preferred AZ to be optimal when in
PlacementPolicy::Secondary
- When transitioning to from PlacementPolicy::Attached(N) to
PlacementPolicy::Secondary, instead of arbitrarily picking a location to
keep, prefer to keep the location in the preferred AZ
We encountered some TLS validation errors for the storcon since applying
#10614. Add an option to downgrade them to logged errors instead to
allow us to debug with more peace.
cc issue https://github.com/neondatabase/cloud/issues/23583
## Problem
* The behavior of this flag changed. Plus, it's not necessary to disable
the IP check as long as there are no IPs listed in the local postgres.
## Summary of changes
* Drop the flag from the command in the README.md section.
* Change the postgres URL passed to proxy to not use the endpoint
hostname.
* Also swap postgres creation and proxy startup, so the DB is running
when proxy comes up.
## Problem
Tests for logical replication (on Staging) have been failing for some
time because logical replication is not enabled for them. This issue
occurred after switching to an org API key with a different default
setting, where logical replication was not enabled by default.
## Summary of changes
- Add `enable_logical_replication` input to
`actions/neon-project-create`
- Enable logical replication in `test-logical-replication` job
Logging errors with the debug format specifier causes multi-line errors,
which are sometimes a pain to deal with. Instead, we should use anyhow's
alternate display format, which shows the same information on a single
line.
Also adjusted a couple of error messages that were stale.
Fixesneondatabase/cloud#14710.
## Problem
1. First of all it's more correct
2. Current usage allows ` Time-of-Check-Time-of-Use (TOCTOU) 'Pwn
Request' vulnerabilities`. Please check security slack channel or reach
me for more details. I will update PR description after merge.
## Summary of changes
1. Use `actions/checkout` with `ref: ${{
github.event.pull_request.head.sha }}`
Discovered by and Co-author: @varunsh-coder
Fixes flaky test_lr_with_slow_safekeeper test #10242
Fix query to `pg_catalog.pg_stat_subscription` catalog to handle table
synchronization and parallel LR correctly.
Successor of #10280 after it was reverted in #10592.
Re-introduce the usage of diesel-async again, but now also add TLS
support so that we connect to the storcon database using TLS. By
default, diesel-async doesn't support TLS, so add some code to make us
explicitly request TLS.
cc https://github.com/neondatabase/cloud/issues/23583
## Problem
While working on https://github.com/neondatabase/neon/pull/10617 I
(unintentionally) merged the PR before the main CI pipeline has
finished.
I suspect this happens because we have received all the required job
results from the pre-merge-checks workflow, which runs on PRs that
include changes to relevant files.
## Summary of changes
- Skip the `conclusion` job in `pre-merge-checks` workflows for PRs
## Problem
This test would sometimes emit unexpected logs from the storage
controller's requests to do migrations, which overlap with the test's
restarts of pageservers, where those migrations are happening some time
after a shard split as the controller moves load around.
Example:
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-10602/13067323736/index.html#testresult/f66f1329557a1fc5/retries
## Summary of changes
- Do a reconcile_until_idle after shard split, so that the rest of the
test doesn't run concurrently with migrations
In the safekeeper, we block deletions on the timeline's gate closing,
and any `WalResidentTimeline` keeps the gate open (because it owns a
gate lock object). Thus, unless the `main_task` function of a partial
backup doesn't return, we can't delete the associated timeline.
In order to make these tasks exit early, we call the cancellation token
of the timeline upon its shutdown. However, the partial backup task
wasn't looking for the cancellation while waiting to acquire a partial
backup permit.
On a staging safekeeper we have been in a situation in the past where
the semaphore was already empty for a duration of many hours, rendering
all attempted deletions unable to proceed until a restart where the
semaphore was reset:
https://neondb.slack.com/archives/C03H1K0PGKH/p1738416586442029
## Problem
when introducing pg17 for job step `Generate matrix for OLAP benchmarks`
I introduced a syntax error that only hits on Saturdays.
## Summary of changes
Remove trailing comma
## successful test run
https://github.com/neondatabase/neon/actions/runs/13086363907
- Wired up filtering on VPC endpoints
- Wired up block access from public internet / VPC depending on per
project flag
- Added cache invalidation for VPC endpoints (partially based on PR from
Raphael)
- Removed BackendIpAllowlist trait
---------
Co-authored-by: Ivan Efremov <ivan@neon.tech>
We forked copy_bidirectional to solve some issues like fast-shutdown
(disallowing half-open connections) and to introduce better error
tracking (which side of the conn closed down).
A change recently made its way upstream offering performance
improvements: https://github.com/tokio-rs/tokio/pull/6532. These seem
applicable to our fork, thus it makes sense to apply them here as well.
The primary benefit is that all the ad hoc get_matches() calls are no
longer necessary. Now all it takes to get at the CLI arguments is
referencing a struct member. It's also great the we can replace the ad
hoc CLI struct we had with this more formal solution.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
Merge Queue fails if changes include Rust code.
## Summary of changes
- Fix condition for `build-build-tools-image`
- Add a couple of no-op `false ||` to make predicates look
symmetric
## Problem
When a client dropped before a request completed, and a handler returned
an ApiError, we would log that at error severity. That was excessive in
the case of a request erroring on a shutdown, and could cause test
flakes.
example:
https://neon-github-public-dev.s3.amazonaws.com/reports/main/13067651123/index.html#suites/ad9c266207b45eafe19909d1020dd987/6021ce86a0d72ae7/
```
Cancelled request finished with an error: ShuttingDown
```
## Summary of changes
- Log a different info-level on ShuttingDown and ResourceUnavailable API
errors from cancelled requests
## Problem
This assertion is incorrect: it is legal to see another shard's data at
this point, after a shard split.
Closes: https://github.com/neondatabase/neon/issues/10609
## Summary of changes
- Remove faulty assertion
## Problem
There are two (related) problems with the previous handling of
`cargo-deny`:
- When a new advisory is added to rustsec that affects a dependency,
unrelated pull requests will fail.
- New advisories rely on pushes or PRs to be surfaced. Problems that
already exist on main will only be found if we try to merge new things
into main.
## Summary of changes
We split out `cargo-deny` into a separate workflow that runs on all PRs
that touch `Cargo.lock`, and on a schedule on `main`, `release`,
`release-compute` and `release-proxy` to find new advisories.
Update to a rebased version of our rust-postgres patches, rebased on
[this](98f5a11bc0)
commit this time.
With #10280 reapplied, this means that the rust-postgres crates will be
deduplicated, as the new crate versions are finally compatible with the
requirements of diesel-async.
Earlier update: #10561
rust-postgres PR: https://github.com/neondatabase/rust-postgres/pull/39
## Problem
We want to check that `diesel print-schema` doesn't generate any changes
(`storage_controller/src/schema.rs`) in comparison with the list of
migration.
## Summary of changes
- Add `diesel_cli` to `build-tools` image
- Add `Check diesel schema` step to `build-neon` job, at this stage we
have all required binaries, so don't need to compile anything
additionally
- Check runs only on x86 release builds to be sure we do it at least
once per CI run.
Ref: https://github.com/neondatabase/cloud/issues/23461
## Problem
Just made changes around and see these 2 base layers could be optimised.
and after review comment from @myrrc setting up timeouts and retries in
`alpine/curl` image
## Summary of changes
## Problem
This test may not fully detect data corruption during splits, since we
don't force-compact the entire keyspace.
## Summary of changes
Force-compact all data in `test_sharding_autosplit`.
## Problem
If offloading races with normal shutdown, we get a "failed to freeze and
flush: cannot flush frozen layers when flush_loop is not running, state
is Exited". This is harmless but points to it being quite strange to try
and freeze and flush such a timeline. flushing on shutdown for an
archived timeline isn't useful.
Related: https://github.com/neondatabase/neon/issues/10389
## Summary of changes
- During Timeline::shutdown, ignore ShutdownMode::FreezeAndFlush if the
timeline is archived
## Problem
test_scrubber_tenant_snapshot restarts pageservers, but log validation
fails tests on any non white listed storcon warnings, making the test
flaky.
## Summary of changes
Allow warns like
2025-01-29T12:37:42.622179Z WARN reconciler{seq=1
tenant_id=2011077aea9b4e8a60e8e8a19407634c shard_id=0004}: Call to node
2 (localhost:15352) management API failed, will retry (attempt 1):
receive body: error sending request for url
(http://localhost:15352/v1/tenant/2011077aea9b4e8a60e8e8a19407634c-0004/location_config):
client error (Connect)
ref https://github.com/neondatabase/neon/issues/10462
Update `tokio` base crates and their deps. Pin `tokio` to at least 1.41
which stabilized task ID APIs.
To dedup `mio` dep the `notify` crate is updated. It's used in
`compute_tools`.
9f81828429/compute_tools/src/pg_helpers.rs (L258-L367)
## Problem
Because dashmap 6 switched to hashbrown RawTable API, it required us to
use unsafe code in the upgrade:
https://github.com/neondatabase/neon/pull/8107
## Summary of changes
Switch to clashmap, a fork maintained by me which removes much of the
unsafe and ultimately switches to HashTable instead of RawTable to
remove much of the unsafe requirement on us.
The test runs this query:
select count(*), sum(data::bigint)::bigint from t
to validate the test results between each part of the test. It performs
a simple sequential scan and aggregation, but was taking an order of
magnitude longer on v17 than on previous Postgres versions, which
sometimes caused the test to time out. There were two reasons for that:
1. On v17, the planner estimates the table to have only only one row. In
reality it has 305790 rows, and older versions estimated it at 611580,
which is not too bad given that the table has not been analyzed so the
planner bases that estimate just on the number of pages and the widths
of the datatypes. The new estimate of 1 row is much worse, and it leads
the planner to disregard parallel plans, whereas on older versions you
got a Parallel Seq Scan.
I tracked this down to upstream commit 29cf61ade3, "Consider fillfactor
when estimating relation size". With that commit,
table_block_relation_estimate_size() function calculates that each page
accommodates less than 1 row when the fillfactor is taken into account,
which rounds down to 0. In reality, the executor will always place at
least one row on a page regardless of fillfactor, but the new estimation
formula doesn't take that into account.
I reported this to pgsql-hackers
(https://www.postgresql.org/message-id/2bf9d973-7789-4937-a7ca-0af9fb49c71e%40iki.fi),
we don't need to do anything more about it in neon. It's OK to not use
parallel scans here; once issue 2. below is addressed, the queries are
fast enough without parallelism..
2. On v17, prefetching was not happening for the sequential scan. That's
because starting with v17, buffers are reserved in the shared buffer
cache before prefetching is initiated, and we use a tiny
shared_buffers=1MB setting in the tests. The prefetching is effectively
disabled with such a small shared_buffers setting, to protect the system
from completely starving out of buffers.
To address that, simply bump up shared_buffers in the test.
This patch addresses the second issue, which is enough to fix the
problem.
## Problem
In https://github.com/neondatabase/neon/pull/10438 it was pointed out
that it would be good to avoid picking tenants in ID order, and also to
avoid situations where we might double-select the same tenant.
There was an initial swing at this in
https://github.com/neondatabase/neon/pull/10443, where Chi suggested a
simpler approach which is done in this PR
## Summary of changes
- Split total set of tenants into in and out of home AZ
- Consume out of home AZ first, and if necessary shuffle + consume from
out of home AZ
## Problem
In https://github.com/neondatabase/neon/pull/10438 I had got the
function for picking tenants backwards, and it was preferring to move
things _away_ from their preferred AZ.
## Summary of changes
- Fix condition in `is_attached_outside_preferred_az`
## Problem
Timeline bootstrap starts a flush loop, but doesn't reliably shut down
the timeline (incl. waiting for flush loop to exit) before destroying
UninitializedTimeline, and that destructor tries to clean up local
storage. If local storage is still being written to, then this is
unsound.
Currently the symptom is that we see a "Directory not empty" error log,
e.g.
https://neon-github-public-dev.s3.amazonaws.com/reports/main/12966756686/index.html#testresult/5523f7d15f46f7f7/retries
## Summary of changes
- Move fallible IO part of bootstrap into a function (notably, this is
fallible in the case of the tenant being shut down while creation is
happening)
- When that function returns an error, call shutdown() on the timeline
## Problem
The test asserts that it completes at least 10 full timeline lifecycles,
but the noisy CI environment sometimes doesn't meet that goal.
Related: https://github.com/neondatabase/neon/issues/10389
## Summary of changes
- Sleep for longer between pageserver restarts, so that the timeline
workers have more chance to make progress
- Sleep for shorter between retries from timeline worker, so that they
have better chance to get in while a pageserver is up between restarts
- Relax the success condition to complete at least 5 iterations instead
of 10
## Problem
for OLAP benchmarks we need specific connstr secrets with different
database names for each job step
This is a follow-up for https://github.com/neondatabase/neon/pull/10536
In previous PR we used a common GitHub secret for a shared re-use
project that has 4 databases: neondb, tpch, clickbench and userexamples.
[Failure
example](https://neon-github-public-dev.s3.amazonaws.com/reports/main/13044872855/index.html#suites/54d0af6f403f1d8611e8894c2e07d023/fc029330265e9f6e/):
```log
# /tmp/neon/pg_install/v17/bin/psql user=neondb_owner dbname=neondb host=ep-broad-brook-w2luwzzv.us-east-2.aws.neon.build sslmode=require options='-cstatement_timeout=0 ' -c -- $ID$
-- TPC-H/TPC-R Pricing Summary Report Query (Q1)
-- Functional Query Definition
-- Approved February 1998
...
ERROR: relation "lineitem" does not exist
```
## Summary of changes
We need dedicated GitHub secrets and dedicated connection strings for
each of the use cases.
## Test run
https://github.com/neondatabase/neon/actions/runs/13053968231
It took me ages to figure out why it was failing on my laptop. What I
saw was that when the test makes the 'import_pgdata' in the pageserver,
the pageserver actually performs a regular 'bootstrap' timeline creation
by running initdb, with no importing. It boiled down to the json request
that the test uses:
```
{
"new_timeline_id": str(timeline_id),
"import_pgdata": {
"idempotency_key": str(idempotency),
"location": {"LocalFs": {"path": str(importbucket.absolute())}},
},
},
```
and how serde deserializes into rust structs. The 'LocalFs' enum variant
in `models.rs` is gated on the 'testing' cargo feature. On a non-testing
build, that got deserialized into the default Bootstrap enum variant, as
a valid TimelineCreateRequestModeImportPgdata variant could not be
formed.
PS. IMHO we should get rid of the testing feature, compile in all the
functionality, and have a runtime flag to disable anything dangeorous.
With that, you would've gotten a nice "feature only enabled in testing
mode" error in this case, or the test would've simply worked. But that's
another story.
## Problem
close https://github.com/neondatabase/neon/issues/10482
## Summary of changes
Add an extra lock on the read path to protect against races. The read
path has an implication that only certain kind of compactions can be
performed. Garbage keys must first have an image layer covering the
range, and then being gc-ed -- they cannot be done in one operation. An
alternative to fix this is to move the layers read guard to be acquired
at the beginning of `get_vectored_reconstruct_data_timeline`, but that
was intentionally optimized out and I don't want to regress.
The race is not limited to image layers. Gc-compaction will consolidate
deltas automatically and produce a flat delta layer (i.e., when we have
retain_lsns below the gc-horizon). The same race would also cause
behaviors like getting an un-replayable key history as in
https://github.com/neondatabase/neon/issues/10049.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
In some cases, we were returning a very shallow error like `error
sending request for url (XXX)`, which made it very hard to figure out
the actual error.
## Summary of changes
Use `{:?}` in a few places, and remove it from places where we were
printing a string anyway.
Related to #10308, we might have legitimate changes in file size or
generation. Those changes should not cause warn log lines.
In order to detect changes of the generation number while the file size
stayed the same, load the metadata that we store on disk on loading of
the timeline.
Still do a comparison with the on-disk layer sizes to find any
discrepancies that might occur due to race conditions (new metadata file
gets written but layer file has not been updated yet, and PS shuts
down). However, as it's possible to hit it in a race conditon, downgrade
it to a warning.
Also fix a mistake in #10529: we want to compare the old with the new
metadata, not the old metadata with itself.
## Problem
The client code for `tenant-set-preferred-az` declared response type
`()`, so printed a spurious error on each use:
```
Error: receive body: error decoding response body: invalid type: map, expected unit at line 1 column 0
```
The requests were successful anyway.
## Summary of changes
- Declare the proper return type, so that the command succeeds quietly.
## Problem
We don't have per-timeline observability for read amplification.
Touches https://github.com/neondatabase/cloud/issues/23283.
## Summary of changes
Add a per-timeline `pageserver_layers_per_read` histogram.
NB: per-timeline histograms are expensive, but probably worth it in this
case.
## Problem
We suspect that Postgres checkpoints will limit the number of page
deltas necessary to reconstruct a page, but don't know for certain.
Touches https://github.com/neondatabase/cloud/issues/23283.
## Summary of changes
Add `pageserver_deltas_per_read_global` metric.
This pairs with `pageserver_layers_per_read_global` from #10573.
## Problem
The current global `pageserver_layers_visited_per_vectored_read_global`
metric does not appear to accurately measure read amplification. It
divides the layer count by the number of reads in a batch, but this
means that e.g. 10 reads with 100 L0 layers will only measure a read amp
of 10 per read, while the actual read amp was 100.
While the cost of layer visits are amortized across the batch, and some
layers may not intersect with a given key, each visited layer
contributes directly to the observed latency for every read in the
batch, which is what we care about.
Touches https://github.com/neondatabase/cloud/issues/23283.
Extracted from #10566.
## Summary of changes
* Count the number of layers visited towards each read in the batch,
instead of the average across the batch.
* Rename `pageserver_layers_visited_per_vectored_read_global` to
`pageserver_layers_per_read_global`.
* Reduce the read amp log warning threshold down from 512 to 100.
## Problem
Follow up of https://github.com/neondatabase/neon/pull/10550 in case the
upper limit is set larger than threshold. It does not make sense for
someone to enforce the behavior like "if there are >= 50 L0s, only
compact 10 of them".
## Summary of changes
Use the maximum of compaction threshold and upper limit when selecting
L0 files to compact.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Repartition is slow, but it's only used in image layer creation. We can
skip it if we have a lot of L0 layers to ingest.
## Summary of changes
If L0 compaction is not complete, do not repartition and do not create
image layers.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We don't have good observability for per-timeline compaction debt,
specifically the number of delta layers in the frozen, L0, and L1
levels.
Touches https://github.com/neondatabase/cloud/issues/23283.
## Summary of changes
* Add a `level` label for `pageserver_layer_{count,size}` with values
`l0`, `l1`, and `frozen`.
* Track metrics for frozen layers.
There is already a `kind={delta,image}` label. `kind=image` is only
possible for `level=l1`.
We don't include the currently open ephemeral layer, only frozen layers.
There is always exactly 1 ephemeral layer, with a dynamic size which is
already tracked in `pageserver_timeline_ephemeral_bytes`.
## Problem
benchmarking.yml so far is only running benchmarks with PostgreSQL
version 16.
However neon recently changed the default for new customers to
PostgreSQL version 17.
See related [epic](https://github.com/neondatabase/cloud/issues/23295)
## Summary of changes
We do not want to run every job step with both pg 16 and 17 because this
would need excessive resources (runners, computes) and extend the
benchmarking run wall clock time too much.
So we select an opinionated subset of testcases that we also report in
weekly reporting and add a postgres v17 job step.
For re-use projects associated Neon projects have been created and
connection strings have been added to neon database organization
secrets.
A follow up is to add the reporting for these new runs to some grafana
dashboards.
## Problem
1. d04d924 added separate metrics for total requests and failures
separately, but it doesn't make much sense. We could just have a unified
counter with `http_status`.
2. `test_compute_migrations_retry` had a race, i.e., it was waiting for
the last successful migration, not an actual failure. This was revealed
after adding an assert on failure metric in d04d924.
## Summary of changes
1. Switch to unified counters for `compute_ctl` requests.
2. Add a waiting loop into `test_compute_migrations_retry` to eliminate
the race.
Part of neondatabase/cloud#17590
## Problem
If we are GC-ing because a new image layer was added while traversing
the timeline, then it will remove layers that are required for
fulfilling the current get request (read-path cannot "look back" and
notice the new image layer).
## Summary of Changes
Prevent GC from progressing on the current timeline while it is being
visited for a read.
Epic: https://github.com/neondatabase/neon/issues/9376
Luckily they were the same version, so we didn't spend time compiling
two versions, which could have been the case in the future.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
The approach of having CancelMap as an in-memory structure increases
code complexity,
as well as putting additional load for Redis streams.
## Summary of changes
- Implement a set of KV ops for Redis client;
- Remove cancel notifications code;
- Send KV ops over the bounded channel to the handling background task
for removing and adding the cancel keys.
Closes#9660
## Problem
We have to test the extensions, shipped with Neon for compatibility
before the upgrade.
## Summary of changes
Added the test for compatibility with the upgraded extensions.
## Problem
Follow-up of the incident, we should not use the same bound on
lower/upper limit of compaction files. This patch adds an upper bound
limit, which is set to 50 for now.
## Summary of changes
Add `compaction_upper_limit`.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
Ref: https://github.com/neondatabase/cloud/issues/23314
We suspect some inconsistency in Benchmark tests runs could be due to
different type of runners they are landed in.
To have that aligned in both terms: failure rates and benchmark results,
lets run them for now on `small-metal` servers and see the progress for
the tests stability.
## Summary of changes
## Problem
There are several parts of `compute_ctl` with a very low visibility of
errors:
1. DB migrations that run async in the background after compute start.
2. Requests made to control plane (currently only `GetSpec`).
3. Requests made to the remote extensions server.
## Summary of changes
Add new counters to quickly evaluate the amount of errors among the
fleet.
Part of neondatabase/cloud#17590
## Problem
https://github.com/neondatabase/neon/pull/10448 removed release notes,
because if their generation failed, the whole release was failing.
People liked them though, and wanted some basic release notes as a
fall-back instead of completely removing them.
## Summary of changes
Include basic release notes that link to the release PR and to a diff to
the previous release.
## Problem
We've seen the ingest connection manager get stuck shortly after a
migration.
## Summary of changes
A speculative mitigation is to use the same mechanism as get page
requests for kicking LSN ingest. The connection manager monitors
LSN waits and queries the broker if no updates are received for the
timeline.
Closes https://github.com/neondatabase/neon/issues/10351
## Problem
We need a setting to disable the flush upload wait, to test L0 flush
backpressure in staging.
## Summary of changes
Add `l0_flush_wait_upload` setting.
## Problem
The request data and usage metrics S3 requests use the same identifier
shown in logs, causing confusion about what type of upload failed.
## Summary of changes
Use the correct identifier for usage metrics uploads.
neondatabase/cloud#23084
Only a few things that needed updating:
- async_trait was removed
- Message::Text takes a Utf8Bytes object instead of a String
Signed-off-by: Tristan Partin <tristan@neon.tech>
Co-authored-by: Conrad Ludgate <connor@neon.tech>
In #10308, we noticed many warnings about the local layer having
different sizes on-disk compared to the metadata.
However, the layer downloader would never redownload layer files if the
sizes or generation numbers change. This is obviously a bug, which we
aim to fix with this PR.
This change also moves the code deciding what to do about a layer to a
dedicated function: before we handled the "routing" via control flow,
but now it's become too complicated and it is nicer to have the
different verdicts for a layer spelled out in a list/match.
This reverts commit 9e55d79803.
We'll still need this until we can tune L0 flush backpressure and
compaction. I'll add a setting to disable this separately.
## Problem
This one is fairly embarrassing. Safekeeper node id was used in the
pageserver application name
when connecting to safekeepers.
## Summary of changes
Use the right node id.
Closes https://github.com/neondatabase/neon/issues/10461
We want to verify if pageserver stripe size has an impact on ingest
performance.
We want to verify if ingest performance has improved or regressed with
postgres version 17.
## Summary of changes
- Allow to create new project with different postgres versions
- allow to pre-shard new project with different stripe sizes instead of
relying on storage manager to shard_split the project once a threshold
is exceeded
Replaces https://github.com/neondatabase/neon/pull/10509
Test run https://github.com/neondatabase/neon/actions/runs/12986410381
We now don't need libpq any more for the build of the storage
controller, as we use `diesel-async` since #10280. Therefore, we remove
the env var that gave cargo/rustc the location for libpq.
Follow-up of #10280
During broker deploys, pageservers log this noisy WARN en masse.
I can trivially reproduce the WARN message in neon_local by SIGKILLing
broker during e.g. `pgbench -i`.
I don't understand why tonic is not detecting the error as
`Code::Unavailable`.
Until we find time to understand that / fix upstream, this PR adds the
error message to the existing list of known error messages that get
demoted to INFO level.
Refs:
- refs https://github.com/neondatabase/neon/issues/9562
## Problem
We were logging a warning after a single request timeout, while listing
objects.
Closes: https://github.com/neondatabase/neon/issues/10166
## Summary of changes
- These timeouts are a pretty normal part of life, so back it off to
only log a warning after two in a row.
## Problem
From time to time, folks discover our `control_plane/` folder and make
the (reasonable) mistake of thinking it's a tool for running full-sized
Neon systems, whereas in reality it is a tool for dev/test.
## Summary of changes
- Change control_plane's readme title to "Local Development Control
Plane (`neon_local`)`
- Change "Running local installation" to "Running a local development
environment" in the main readme
## Problem
The intent of this parameter is to have pageservers consider themselves
"full" if they've got lots of shards, even if they have plenty of
capacity. It works, but because we typically successfully oversubscribe
capacity up to 200%, the MAX_SHARDS limit is effectively doubled, so
this 20,000 value ends up meaning 40,000, whereas the original intent
was to limit nodes to ~10000 shards.
## Summary of changes
- Change MAX_SHARDS to 5000, so that a node with 5000 will get a 100%
utilization, which is equivalent in practice to being considered "half
full" by the storage controller in capacity terms.
This is all a bit subtle and indiret. Originally the limit was baked
into the pageserver with the idea that the pageserver knows better what
its own resources tolerate than the storage controller does, but in
practice it would be probably be easier to understand all this if we
just did it controller-side. So there's scope to refactor here in
future.
Switches the storcon away from using diesel's synchronous APIs in favour
of `diesel-async`.
Advantages:
* less C dependencies, especially no openssl, which might be behind the
bug: https://github.com/neondatabase/cloud/issues/21010
* Better to only have async than mix of async plus `spawn_blocking`
We had to turn off usage of the connection pool for migrations, as
diesel migrations don't support async APIs. Thus we still use
`spawn_blocking` in that one place. But this is explicitly done in one
of the `diesel-async` examples.
## Problem
In ingest benchmarks, we see L0 compaction delays of over 10 minutes due
to image compaction. We can't stall L0 flushes for that long.
## Summary of changes
Disable L0 flush stalls, and bump the default L0 flush delay threshold
from 20 to 30 L0 layers.
## Problem
If compaction fails, we disable L0 flush stalls to avoid persistent
stalls. However, the logic would unset the failure marker on offload
failures or shutdown. This can lead to sudden L0 flush stalls if we try
and fail to offload a timeline with compaction failures, or if there is
some kind of shutdown race.
Touches #10405.
## Summary of changes
Don't touch the compaction failure marker on offload failures or
shutdown.
## Problem
After talking about it again with @bayandin again this should replace
the changes from https://github.com/neondatabase/neon/pull/10475. While
the previous changes worked, they are less visually clear in what
happens, and we might end up in a situation where we update `latest`,
but don't actually have the tagged image pushed that contains the same
changes. The latter would result in potentially hard to debug
situations.
## Summary of changes
Revert c283aaaf8d and make
promote-images-prod depend on promote-images-dev instead.
## Problem
The containers' log output is mixed with the tests' output, so you must
scroll up to find the error.
## Summary of changes
Printing of containers' logs moved to a separate step.
Note: this has to merge after the release is cut on `2025-01-17` for
compat tests to start passing.
## Problem
SK wal reader fan-out is not enabled in tests by default.
## Summary of changes
Enable it.
## Problem
There is no direct backpressure for compaction and L0 read
amplification. This allows a large buildup of compaction debt and read
amplification.
Resolves#5415.
Requires #10402.
## Summary of changes
Delay layer flushes based on the number of level 0 delta layers:
* `l0_flush_delay_threshold`: delay flushes such that they take 2x as
long (default `2 * compaction_threshold`).
* `l0_flush_stall_threshold`: stall flushes until level 0 delta layers
drop below threshold (default `4 * compaction_threshold`).
If either threshold is reached, ephemeral layer rolls also synchronously
wait for layer flushes to propagate this backpressure up into WAL
ingestion. This will bound the number of frozen layers to 1 once
backpressure kicks in, since all other frozen layers must flush before
the rolled layer.
## Analysis
This will significantly change the compute backpressure characteristics.
Recall the three compute backpressure knobs:
* `max_replication_write_lag`: 500 MB (based on Pageserver
`last_received_lsn`).
* `max_replication_flush_lag`: 10 GB (based on Pageserver
`disk_consistent_lsn`).
* `max_replication_apply_lag`: disabled (based on Pageserver
`remote_consistent_lsn`).
Previously, the Pageserver would keep ingesting WAL and build up
ephemeral layers and L0 layers until the compute hit
`max_replication_flush_lag` at 10 GB and began backpressuring. Now, once
we delay/stall WAL ingestion, the compute will begin backpressuring
after `max_replication_write_lag`, i.e. 500 MB. This is probably a good
thing (we're not building up a ton of compaction debt), but we should
consider tuning these settings.
`max_replication_flush_lag` probably doesn't serve a purpose anymore,
and we should consider removing it.
Furthermore, the removal of the upload barrier in #10402 will mean that
we no longer backpressure flushes based on S3 uploads, since
`max_replication_apply_lag` is disabled. We should consider enabling
this as well.
### When and what do we compact?
Default compaction settings:
* `compaction_threshold`: 10 L0 delta layers.
* `compaction_period`: 20 seconds (between each compaction loop check).
* `checkpoint_distance`: 256 MB (size of L0 delta layers).
* `l0_flush_delay_threshold`: 20 L0 delta layers.
* `l0_flush_stall_threshold`: 40 L0 delta layers.
Compaction characteristics:
* Minimum compaction volume: 10 layers * 256 MB = 2.5 GB.
* Additional compaction volume (assuming 128 MB/s WAL): 128 MB/s * 20
seconds = 2.5 GB (10 L0 layers).
* Required compaction bandwidth: 5.0 GB / 20 seconds = 256 MB/s.
### When do we hit `max_replication_write_lag`?
Depending on how fast compaction and flushes happens, the compute will
backpressure somewhere between `l0_flush_delay_threshold` or
`l0_flush_stall_threshold` + `max_replication_write_lag`.
* Minimum compute backpressure lag: 20 layers * 256 MB + 500 MB = 5.6 GB
* Maximum compute backpressure lag: 40 layers * 256 MB + 500 MB = 10.0
GB
This seems like a reasonable range to me.
This reapplies #10135. Just removing this flush backpressure without
further mitigations caused read amp increases during bulk ingestion
(predictably), so it was reverted. We will replace it by
compaction-based backpressure.
## Problem
In #8550, we made the flush loop wait for uploads after every layer.
This was to avoid unbounded buildup of uploads, and to reduce compaction
debt. However, the approach has several problems:
* It prevents upload parallelism.
* It prevents flush and upload pipelining.
* It slows down ingestion even when there is no need to backpressure.
* It does not directly backpressure based on compaction debt and read
amplification.
We will instead implement compaction-based backpressure in a PR
immediately following this removal (#5415).
Touches #5415.
Touches #10095.
## Summary of changes
Remove waiting on the upload queue in the flush loop.
## Problem
If gc-compaction decides to rewrite an image layer, it will now cause
index_part to lose reference to that layer. In details,
* Assume there's only one image layer of key 0000...AAAA at LSN 0x100
and generation 0xA in the system.
* gc-compaction kicks in at gc-horizon 0x100, and then produce
0000...AAAA at LSN 0x100 and generation 0xB.
* It submits a compaction result update into the index part that unlinks
0000-AAAA-100-A and adds 0000-AAAA-100-B
On the remote storage / local disk side, this is fine -- it unlinks
things correctly and uploads the new file. However, the
`index_part.json` itself doesn't record generations. The buggy procedure
is as follows:
1. upload the new file
2. update the index part to remove the old file and add the new file
3. remove the new file
Therefore, the correct update result process for gc-compaction should be
as follows:
* When modifying the layer map, delete the old one and upload the new
one.
* When updating the index, uploading the new one in the index without
deleting the old one.
## Summary of changes
* Modify `finish_gc_compaction` to correctly order insertions and
deletions.
* Update the way gc-compaction uploads the layer files.
* Add new tests.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
We've finally transitioned to using a separate `release-compute` branch.
Now, we can finally automatically create release PRs on Fri and release
them during the following week.
Part of neondatabase/cloud#11698
Sometimes, especially when the host running the tests is overloaded, we
can run into reconcile timeouts in
`test_timeline_ancestor_detach_idempotent_success`, making the test
flaky. By increasing the timeouts from 30 seconds to 120 seconds, we can
address the flakiness.
Fixes#10464
## Problem
Currently, the report does not contain the LFC state of the failed
tests.
## Summary of changes
Added the LFC state to the link to the allure report.
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
Drop logical replication subscribers
before compute starts on a non-main branch.
Add new compute_ctl spec flag: drop_subscriptions_before_start
If it is set, drop all the subscriptions from the compute node
before it starts.
To avoid race on compute start, use new GUC
neon.disable_logical_replication_subscribers
to temporarily disable logical replication workers until we drop the
subscriptions.
Ensure that we drop subscriptions exactly once when endpoint starts on a
new branch.
It is essential, because otherwise, we may drop not only inherited, but
newly created subscriptions.
We cannot rely only on spec.drop_subscriptions_before_start flag,
because if for some reason compute restarts inside VM,
it will start again with the same spec and flag value.
To handle this, we save the fact of the operation in the database
in the neon.drop_subscriptions_done table.
If the table does not exist, we assume that the operation was never
performed, so we must do it.
If table exists, we check if the operation was performed on the current
timeline.
fixes: https://github.com/neondatabase/neon/issues/8790
## Problem
Not really a bug fix, but hopefully can reproduce
https://github.com/neondatabase/neon/issues/10482 more.
If the layer map does not contain layers that end at exactly the end
range of the compaction job, the current split algorithm will produce
the last job that ends at the maximum layer key. This patch extends it
all the way to the compaction job end key.
For example, the user requests a compaction of 0000...FFFF. However, we
only have a layer 0000..3000 in the layer map, and the split job will
have a range of 0000..3000 instead of 0000..FFFF.
This is not a correctness issue but it would be better to fix it so that
we can get consistent job splits.
## Summary of changes
Compaction job split will always cover the full specified key range.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
PR #10457 was supposed to fix the flakiness of
`test_scrubber_physical_gc_ancestors`, but instead it made it even more
flaky. However, the original error causes disappeared, now to be
replaced by key not found errors.
See this for a longer explanation:
https://github.com/neondatabase/neon/issues/10391#issuecomment-2608018967
## Solution
This does one churn rows after all compactions, and before we do any
timeline gc's. That way, we remain more accessible at older lsn's.
## Problem
The trust connection to the compute required for `pg_anon` was removed.
However, the PGPASSWORD environment variable was not added to
`docker-compose.yml`.
This caused connection errors, which were interpreted as success due to
errors in the bash script.
## Summary of changes
The environment variable was added, and the logic in the bash script was
fixed.
## Problem
https://github.com/neondatabase/neon/actions/runs/12896686483/job/35961290336#step:5:107
showed that `promote-images-prod` was missing another dependency.
## Summary of changes
Modify `promote-images-prod` to tag based on docker-hub images, so that
`promote-images-prod` does not rely on `promote-images-dev`. The result
should be the exact same, but allows the two jobs to run in parallel.
## Refs
- Epic: https://github.com/neondatabase/neon/issues/9378
Co-authored-by: Vlad Lazar <vlad@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
The read path does its IOs sequentially.
This means that if N values need to be read to reconstruct a page,
we will do N IOs and getpage latency is `O(N*IoLatency)`.
## Solution
With this PR we gain the ability to issue IO concurrently within one
layer visit **and** to move on to the next layer without waiting for IOs
from the previous visit to complete.
This is an evolved version of the work done at the Lisbon hackathon,
cf https://github.com/neondatabase/neon/pull/9002.
## Design
### `will_init` now sourced from disk btree index keys
On the algorithmic level, the only change is that the
`get_values_reconstruct_data`
now sources `will_init` from the disk btree index key (which is
PS-page_cache'd), instead
of from the `Value`, which is only available after the IO completes.
### Concurrent IOs, Submission & Completion
To separate IO submission from waiting for its completion, while
simultaneously
feature-gating the change, we introduce the notion of an `IoConcurrency`
struct
through which IO futures are "spawned".
An IO is an opaque future, and waiting for completions is handled
through
`tokio::sync::oneshot` channels.
The oneshot Receiver's take the place of the `img` and `records` fields
inside `VectoredValueReconstructState`.
When we're done visiting all the layers and submitting all the IOs along
the way
we concurrently `collect_pending_ios` for each value, which means
for each value there is a future that awaits all the oneshot receivers
and then calls into walredo to reconstruct the page image.
Walredo is now invoked concurrently for each value instead of
sequentially.
Walredo itself remains unchanged.
The spawned IO futures are driven to completion by a sidecar tokio task
that
is separate from the task that performs all the layer visiting and
spawning of IOs.
That tasks receives the IO futures via an unbounded mpsc channel and
drives them to completion inside a `FuturedUnordered`.
(The behavior from before this PR is available through
`IoConcurrency::Sequential`,
which awaits the IO futures in place, without "spawning" or "submitting"
them
anywhere.)
#### Alternatives Explored
A few words on the rationale behind having a sidecar *task* and what
alternatives were considered.
One option is to queue up all IO futures in a FuturesUnordered that is
polled
the first time when we `collect_pending_ios`.
Firstly, the IO futures are opaque, compiler-generated futures that need
to be polled at least once to submit their IO. "At least once" because
tokio-epoll-uring may not be able to submit the IO to the kernel on
first
poll right away.
Second, there are deadlocks if we don't drive the IO futures to
completion
independently of the spawning task.
The reason is that both the IO futures and the spawning task may hold
some
_and_ try to acquire _more_ shared limited resources.
For example, both spawning task and IO future may try to acquire
* a VirtualFile file descriptor cache slot async mutex (observed during
impl)
* a tokio-epoll-uring submission slot (observed during impl)
* a PageCache slot (currently this is not the case but we may move more
code into the IO futures in the future)
Another option is to spawn a short-lived `tokio::task` for each IO
future.
We implemented and benchmarked it during development, but found little
throughput improvement and moderate mean & tail latency degradation.
Concerns about pressure on the tokio scheduler made us discard this
variant.
The sidecar task could be obsoleted if the IOs were not arbitrary code
but a well-defined struct.
However,
1. the opaque futures approach taken in this PR allows leaving the
existing
code unchanged, which
2. allows us to implement the `IoConcurrency::Sequential` mode for
feature-gating
the change.
Once the new mode sidecar task implementation is rolled out everywhere,
and `::Sequential` removed, we can think about a descriptive submission
& completion interface.
The problems around deadlocks pointed out earlier will need to be solved
then.
For example, we could eliminate VirtualFile file descriptor cache and
tokio-epoll-uring slots.
The latter has been drafted in
https://github.com/neondatabase/tokio-epoll-uring/pull/63.
See the lengthy doc comment on `spawn_io()` for more details.
### Error handling
There are two error classes during reconstruct data retrieval:
* traversal errors: index lookup, move to next layer, and the like
* value read IO errors
A traversal error fails the entire get_vectored request, as before this
PR.
A value read error only fails that value.
In any case, we preserve the existing behavior that once
`get_vectored` returns, all IOs are done. Panics and failing
to poll `get_vectored` to completion will leave the IOs dangling,
which is safe but shouldn't happen, and so, a rate-limited
log statement will be emitted at warning level.
There is a doc comment on `collect_pending_ios` giving more code-level
details and rationale.
### Feature Gating
The new behavior is opt-in via pageserver config.
The `Sequential` mode is the default.
The only significant change in `Sequential` mode compared to before
this PR is the buffering of results in the `oneshot`s.
## Code-Level Changes
Prep work:
* Make `GateGuard` clonable.
Core Feature:
* Traversal code: track `will_init` in `BlobMeta` and source it from
the Delta/Image/InMemory layer index, instead of determining `will_init`
after we've read the value. This avoids having to read the value to
determine whether traversal can stop.
* Introduce `IoConcurrency` & its sidecar task.
* `IoConcurrency` is the clonable handle.
* It connects to the sidecar task via an `mpsc`.
* Plumb through `IoConcurrency` from high level code to the
individual layer implementations' `get_values_reconstruct_data`.
We piggy-back on the `ValuesReconstructState` for this.
* The sidecar task should be long-lived, so, `IoConcurrency` needs
to be rooted up "high" in the call stack.
* Roots as of this PR:
* `page_service`: outside of pagestream loop
* `create_image_layers`: when it is called
* `basebackup`(only auxfiles + replorigin + SLRU segments)
* Code with no roots that uses `IoConcurrency::sequential`
* any `Timeline::get` call
* `collect_keyspace` is a good example
* follow-up: https://github.com/neondatabase/neon/issues/10460
* `TimelineAdaptor` code used by the compaction simulator, unused in
practive
* `ingest_xlog_dbase_create`
* Transform Delta/Image/InMemoryLayer to
* do their values IO in a distinct `async {}` block
* extend the residence of the Delta/Image layer until the IO is done
* buffer their results in a `oneshot` channel instead of straight
in `ValuesReconstructState`
* the `oneshot` channel is wrapped in `OnDiskValueIo` /
`OnDiskValueIoWaiter`
types that aid in expressiveness and are used to keep track of
in-flight IOs so we can print warnings if we leave them dangling.
* Change `ValuesReconstructState` to hold the receiving end of the
`oneshot` channel aka `OnDiskValueIoWaiter`.
* Change `get_vectored_impl` to `collect_pending_ios` and issue walredo
concurrently, in a `FuturesUnordered`.
Testing / Benchmarking:
* Support queue-depth in pagebench for manual benchmarkinng.
* Add test suite support for setting concurrency mode ps config
field via a) an env var and b) via NeonEnvBuilder.
* Hacky helper to have sidecar-based IoConcurrency in tests.
This will be cleaned up later.
More benchmarking will happen post-merge in nightly benchmarks, plus in
staging/pre-prod.
Some intermediate helpers for manual benchmarking have been preserved in
https://github.com/neondatabase/neon/pull/10466 and will be landed in
later PRs.
(L0 layer stack generator!)
Drive-By:
* test suite actually didn't enable batching by default because
`config.compatibility_neon_binpath` is always Truthy in our CI
environment
=> https://neondb.slack.com/archives/C059ZC138NR/p1737490501941309
* initial logical size calculation wasn't always polled to completion,
which was
surfaced through the added WARN logs emitted when dropping a
`ValuesReconstructState` that still has inflight IOs.
* remove the timing histograms
`pageserver_getpage_get_reconstruct_data_seconds`
and `pageserver_getpage_reconstruct_seconds` because with planning,
value read
IO, and walredo happening concurrently, one can no longer attribute
latency
to any one of them; we'll revisit this when Vlad's work on
tracing/sampling
through RequestContext lands.
* remove code related to `get_cached_lsn()`.
The logic around this has been dead at runtime for a long time,
ever since the removal of the materialized page cache in #8105.
## Testing
Unit tests use the sidecar task by default and run both modes in CI.
Python regression tests and benchmarks also use the sidecar task by
default.
We'll test more in staging and possibly preprod.
# Future Work
Please refer to the parent epic for the full plan.
The next step will be to fold the plumbing of IoConcurrency
into RequestContext so that the function signatures get cleaned up.
Once `Sequential` isn't used anymore, we can take the next
big leap which is replacing the opaque IOs with structs
that have well-defined semantics.
---------
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
Both these versions are binary compatible, but the way pgvector
structures the SQL files forbids installing 0.7.4 if you have a 0.8.0
distribution. Yet, some users may need a previous version for backward
compatibility, e.g., restoring the dump.
See this thread for discussion
https://neondb.slack.com/archives/C04DGM6SMTM/p1735911490242919?thread_ts=1731343604.259169&cid=C04DGM6SMTM
## Summary of changes
Put `vector--0.7.4.sql` file into compute image to allow installing this
version as well.
Tested on staging and it seems to be working as expected:
```sql
select * from pg_available_extensions where name = 'vector';
name | default_version | installed_version | comment
--------+-----------------+-------------------+------------------------------------------------------
vector | 0.8.0 | (null) | vector data type and ivfflat and hnsw access methods
create extension vector version '0.7.4';
select * from pg_available_extensions where name = 'vector';
name | default_version | installed_version | comment
--------+-----------------+-------------------+------------------------------------------------------
vector | 0.8.0 | 0.7.4 | vector data type and ivfflat and hnsw access methods
alter extension vector update;
select * from pg_available_extensions where name = 'vector';
name | default_version | installed_version | comment
--------+-----------------+-------------------+------------------------------------------------------
vector | 0.8.0 | 0.8.0 | vector data type and ivfflat and hnsw access methods
drop extension vector;
create extension vector;
select * from pg_available_extensions where name = 'vector';
name | default_version | installed_version | comment
--------+-----------------+-------------------+------------------------------------------------------
vector | 0.8.0 | 0.8.0 | vector data type and ivfflat and hnsw access methods
```
If we find out it's a good approach, we can adopt the same for other
extensions with a stable ABI -- support both `current` and `current - 1`
releases.
# Refs
- extracted from https://github.com/neondatabase/neon/pull/9353
# Problem
Before this PR, when task_mgr shutdown is signalled, e.g. during
pageserver shutdown or Tenant shutdown, initial logical size calculation
stops polling and drops the future that represents the calculation.
This is against the current policy that we poll all futures to
completion.
This became apparent during development of concurrent IO which warns if
we drop a `Timeline::get_vectored` future that still has in-flight IOs.
We may revise the policy in the future, but, right now initial logical
size calculation is the only part of the codebase that doesn't adhere to
the policy, so let's fix it.
## Code Changes
- make sensitive exclusively to `Timeline::cancel`
- This should be sufficient for all cases of shutdowns; the sensitivity
to task_mgr shutdown is unnecessary.
- this broke the various cancel tests in `test_timeline_size.py`, e.g.,
`test_timeline_initial_logical_size_calculation_cancellation`
- the tests would time out because the await point was not sensitive to
cancellation
- to fix this, refactor `pausable_failpoint` so that it accepts a
cancellation token
- side note: we _really_ should write our own failpoint library; maybe
after we get heap-allocated RequestContext, we can plumb failpoints
through there.
## Problem
PR #9993 was supposed to enable `page_service_pipelining` by default for
all `NeonEnv`s, but this was ineffective in our CI environment.
Thus, CI Python-based tests and benchmarks, unless explicitly
configuring pipelining, were still using serial protocol handling.
## Analysis
The root cause was that in our CI environment,
`config.compatibility_neon_binpath` is always Truthy.
It's not in local environments, which is why this slipped through in
local testing.
Lesson: always add a log line ot pageserver startup and spot-check tests
to ensure the intended default is picked up.
## Summary of changes
Fix it. Since enough time has passed, the compatiblity snapshot contains
a recent enough software version so we don't need to worry about
`compatibility_neon_binpath` anymore.
## Future Work
The question how to add a new default except for compatibliity tests,
which is what the broken code was supposed to do, is still unsolved.
Slack discussion:
https://neondb.slack.com/archives/C059ZC138NR/p1737490501941309
## Problem
Currently, the layer flush loop will continue flushing layers as long as
any are pending, and only notify waiters once there are no further
layers to flush. This can cause waiters to wait longer than necessary,
and potentially starve them if pending layers keep arriving faster than
they can be flushed. The impact of this will increase when we add
compaction backpressure and propagate it up into the WAL receiver.
Extracted from #10405.
## Summary of changes
Break out of the layer flush loop once we've flushed up to the requested
LSN. If further flush requests have arrived in the meanwhile, flushing
will resume immediately after.
## Problem
For compaction backpressure, we need a mechanism to signal when
compaction has reduced the L0 delta layer count below the backpressure
threshold.
Extracted from #10405.
## Summary of changes
Add `LayerMap::watch_level0_deltas()` which returns a
`tokio::sync:⌚:Receiver` signalling the current L0 delta layer
count.
## Problem
It's sometimes useful to obtain the elapsed duration from a
`StorageTimeMetricsTimer` for purposes beyond just recording it in
metrics (e.g. to log it).
Extracted from #10405.
## Summary of changes
Add `StorageTimeMetricsTimer.elapsed()` and return the duration from
`stop_and_record()`.
## Problem
part of https://github.com/neondatabase/neon/issues/9114
The automatic trigger is already implemented at
https://github.com/neondatabase/neon/pull/10221 but I need to write some
tests and finish my experiments in staging before I can merge it with
confidence. Given that I have some other patches that will modify the
config items, I'd like to get the config items merged first to reduce
conflicts.
## Summary of changes
* add `l2_lsn` to index_part.json -- below that LSN, data have been
processed by gc-compaction
* add a set of gc-compaction auto trigger control items into the config
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We are removing the `pg_anon` v1 extension from Neon. So we don't need
to test it anymore and can remove the code for simplicity.
## Summary of changes
The code required for testing `pg_anon` is removed.
We did not have any tests on fast_import binary yet.
In this PR I have introduced:
- `FastImport` class and tools for testing in python
- basic test that runs fast import against vanilla postgres and checks
that data is there
Should be merged after https://github.com/neondatabase/neon/pull/10251
We currently have some flakiness in
`test_scrubber_physical_gc_ancestors`, see #10391.
The first flakiness kind is about the reconciler not actually becoming
idle within the timeout of 30 seconds. We see continuous forward
progress so this is likely not a hang. We also see this happen in
parallel to a test failure, so is likely due to runners being
overloaded. Therefore, we increase the timeout.
The second flakiness kind is an assertion failure. This one is a little
bit more tricky, but we saw in the successful run that there was some
advance of the lsn between the compaction ran (which created layer
files) and the gc run. Apparently gc rejects reductions to the single
image layer setting if the cutoff lsn is the same as the lsn of the
image layer: it will claim that that layer is newer than the space
cutoff and therefore skip it, while thinking the old layer (that we want
to delete) is the latest one (so it's not deleted).
We address the second flakiness kind by inserting a tiny amount of WAL
between the compaction and gc. This should hopefully fix things.
Related issue: #10391
(not closing it with the merger of the PR as we'll need to validate that
these changes had the intended effect).
Thanks to Chi for going over this together with me in a call.
## Problem
When releasing `release-7574`, the Github Release creation failed with
"body is too long" (see
https://github.com/neondatabase/neon/actions/runs/12834025431/job/35792346745#step:5:77).
There's lots of room for improvement of the release notes, but for now
we'll disable them instead.
## Summary of changes
- Disable automatic generation of release notes for Github releases
- Enable creation of Github releases for proxy/compute
This simplifies the code in `pageserver_physical_gc` a little bit after
the feedback in #10007 that the code is too complicated.
Most importantly, we don't pass around `GcSummary` any more in a
complicated fashion, and we save on async stream-combinator-inception in
one place in favour of `try_stream!{}`.
Follow-up of #10007
Otherwise we might hit ERRORs in otherwise safe situations (such as user
queries), which isn't a great user experience.
## Problem
https://github.com/neondatabase/neon/pull/10376
## Summary of changes
Instead of accepting internal errors as acceptable, we ensure we don't
exceed our allocated usage.
## Refs
- fixes https://github.com/neondatabase/neon/issues/10444
## Problem
We're seeing a panic `handles are only shut down once in their lifetime`
in our performance testbed.
## Hypothesis
Annotated code in
https://github.com/neondatabase/neon/issues/10444#issuecomment-2602286415.
```
T1: drop Cache, executes up to (1)
=> HandleInner is now in state ShutDown
T2: Timeline::shutdown => PerTimelineState::shutdown executes shutdown() again => panics
```
Likely this snuck in the final touches of #10386 where I narrowed down
the locking rules.
## Summary of changes
Make duplicate shutdowns a no-op.
## Summary
Whereas currently we send all WAL to all pageserver shards, and each
shard filters out the data that it needs,
in this RFC we add a mechanism to filter the WAL on the safekeeper, so
that each shard receives
only the data it needs.
This will place some extra CPU load on the safekeepers, in exchange for
reducing the network bandwidth
for ingesting WAL back to scaling as O(1) with shard count, rather than
O(N_shards).
Touches #9329.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Vlad Lazar <vlalazar.vlad@gmail.com>
Co-authored-by: Vlad Lazar <vlad@neon.tech>
The other test focus on the external interface usage while the tests
added in this PR add some testing around HandleInner's lifecycle,
ensuring we don't leak it once either connection gets dropped or
per-timeline-state is shut down explicitly.
It was requested by review in #10305 to use an enum or something like it
for distinguishing the different modes instead of two parameters,
because two flags allow four combinations, and two of them don't really
make sense/ aren't used.
follow-up of #10305
## Problem
Since #9916 , the chaos code is actively fighting the optimizer: tenants
tend to be attached in their preferred AZ, so most chaos migrations were
moving them to a non-preferred AZ.
## Summary of changes
- When picking migrations, prefer to migrate things _toward_ their
preferred AZ when possible. Then pick shards to move the other way when
necessary.
The resulting behavior should be an alternating "back and forth" where
the chaos code migrates thiings away from home, and then migrates them
back on the next iteration.
The side effect will be that the chaos code actively helps to push
things into their home AZ. That's not contrary to its purpose though: we
mainly just want it to continuously migrate things to exercise
migration+notification code.
## Problem
Occasionally, we encounter bugs in test environments that can be
detected at the point of uploading an index, but we proceed to upload it
anyway and leave a tenant in a broken state that's awkward to handle.
## Summary of changes
- Validate index when submitting it for upload, so that we can see the
issue quickly e.g. in an API invoking compaction
- Validate index before executing the upload, so that we have a hard
enforcement that any code path that tries to upload an index will not
overwrite a valid index with an invalid one.
Add an endpoint to obtain the utilization of a safekeeper. Future
changes to the storage controller can use this endpoint to find the most
suitable safekeepers for newly created timelines, analogously to how
it's done for pageservers already.
Initially we just want to assign by timeline count, then we can iterate
from there.
Part of https://github.com/neondatabase/neon/issues/9011
## Problem
871e8b325f failed CI on main because a job
ran to soon. This was caused by
ea84ec357f. While `promote-images-dev`
does not inherently need `neon-image`, a few jobs depending on
`promote-images-dev` do need it, and previously had it when it was
`promote-images`, which depended on `test-images`, which in turn
depended on `neon-image`.
## Summary of changes
To ensure jobs depending `docker.io/neondatabase/neon` images get them,
`promote-images-dev` gets the dependency to `neon-image` back which it
previously had transitively through `test-images`.
Instead of generating our own request ID, we can just use the one
provided by the control plane. In the event, we get a request from a
client which doesn't set X-Request-ID, then we just generate one which
is useful for tracing purposes.
Signed-off-by: Tristan Partin <tristan@neon.tech>
# Refs
- fixes https://github.com/neondatabase/neon/issues/10309
- fixup of batching design, first introduced in
https://github.com/neondatabase/neon/pull/9851
- refinement of https://github.com/neondatabase/neon/pull/8339
# Problem
`Tenant::shutdown` was occasionally taking many minutes (sometimes up to
20) in staging and prod if the
`page_service_pipelining.mode="concurrent-futures"` is enabled.
# Symptoms
The issue happens during shard migration between pageservers.
There is page_service unavailability and hence effectively downtime for
customers in the following case:
1. The source (state `AttachedStale`) gets stuck in `Tenant::shutdown`,
waiting for the gate to close.
2. Cplane/Storcon decides to transition the target `AttachedMulti` to
`AttachedSingle`.
3. That transition comes with a bump of the generation number, causing
the `PUT .../location_config` endpoint to do a full `Tenant::shutdown` /
`Tenant::attach` cycle for the target location.
4. That `Tenant::shutdown` on the target gets stuck, waiting for the
gate to close.
5. Eventually the gate closes (`close completed`), correlating with a
`page_service` connection handler logging that it's exiting because of a
network error (`Connection reset by peer` or `Broken pipe`).
While in (4):
- `Tenant::shutdown` is stuck waiting for all `Timeline::shutdown` calls
to complete.
So, really, this is a `Timeline::shutdown` bug.
- retries from Cplane/Storcon to complete above state transitions, fail
with errors related to the tenant mgr slot being in state
`TenantSlot::InProgress`, the tenant state being
`TenantState::Stopping`, and the timelines being in
`TimelineState::Stopping`, and the `Timeline::cancel` being cancelled.
- Existing (and/or new?) page_service connections log errors `error
reading relation or page version: Not found: Timed out waiting 30s for
tenant active state. Latest state: None`
# Root-Cause
After a lengthy investigation ([internal
write-up](https://www.notion.so/neondatabase/2025-01-09-batching-deadlock-Slow-Log-Analysis-in-Staging-176f189e00478050bc21c1a072157ca4?pvs=4))
I arrived at the following root cause.
The `spsc_fold` channel (`batch_tx`/`batch_rx`) that connects the
Batcher and Executor stages of the pipelined mode was storing a `Handle`
and thus `GateGuard` of the Timeline that was not shutting down.
The design assumption with pipelining was that this would always be a
short transient state.
However, that was incorrect: the Executor was stuck on writing/flushing
an earlier response into the connection to the client, i.e., socket
write being slow because of TCP backpressure.
The probable scenario of how we end up in that case:
1. Compute backend process sends a continuous stream of getpage prefetch
requests into the connection, but never reads the responses (why this
happens: see Appendix section).
2. Batch N is processed by Batcher and Executor, up to the point where
Executor starts flushing the response.
3. Batch N+1 is procssed by Batcher and queued in the `spsc_fold`.
4. Executor is still waiting for batch N flush to finish.
5. Batcher eventually hits the `TimeoutReader` error (10min).
From here on it waits on the
`spsc_fold.send(Err(QueryError(TimeoutReader_error)))`
which will never finish because the batch already inside the `spsc_fold`
is not
being read by the Executor, because the Executor is still stuck in the
flush.
(This state is not observable at our default `info` log level)
6. Eventually, Compute backend process is killed (`close()` on the
socket) or Compute as a whole gets killed (probably no clean TCP
shutdown happening in that case).
7. Eventually, Pageserver TCP stack learns about (6) through RST packets
and the Executor's flush() call fails with an error.
8. The Executor exits, dropping `cancel_batcher` and its end of the
spsc_fold.
This wakes Batcher, causing the `spsc_fold.send` to fail.
Batcher exits.
The pipeline shuts down as intended.
We return from `process_query` and log the `Connection reset by peer` or
`Broken pipe` error.
The following diagram visualizes the wait-for graph at (5)
```mermaid
flowchart TD
Batcher --spsc_fold.send(TimeoutReader_error)--> Executor
Executor --flush batch N responses--> socket.write_end
socket.write_end --wait for TCP window to move forward--> Compute
```
# Analysis
By holding the GateGuard inside the `spsc_fold` open, the pipelining
implementation
violated the principle established in
(https://github.com/neondatabase/neon/pull/8339).
That is, that `Handle`s must only be held across an await point if that
await point
is sensitive to the `<Handle as Deref<Target=Timeline>>::cancel` token.
In this case, we were holding the Handle inside the `spsc_fold` while
awaiting the
`pgb_writer.flush()` future.
One may jump to the conclusion that we should simply peek into the
spsc_fold to get
that Timeline cancel token and be sensitive to it during flush, then.
But that violates another principle of the design from
https://github.com/neondatabase/neon/pull/8339.
That is, that the page_service connection lifecycle and the Timeline
lifecycles must be completely decoupled.
Tt must be possible to shut down one shard without shutting down the
page_service connection, because on that single connection we might be
serving other shards attached to this pageserver.
(The current compute client opens separate connections per shard, but,
there are plans to change that.)
# Solution
This PR adds a `handle::WeakHandle` struct that does _not_ hold the
timeline gate open.
It must be `upgrade()`d to get a `handle::Handle`.
That `handle::Handle` _does_ hold the timeline gate open.
The batch queued inside the `spsc_fold` only holds a `WeakHandle`.
We only upgrade it while calling into the various `handle_` methods,
i.e., while interacting with the `Timeline` via `<Handle as
Deref<Target=Timeline>>`.
All that code has always been required to be (and is!) sensitive to
`Timeline::cancel`, and therefore we're guaranteed to bail from it
quickly when `Timeline::shutdown` starts.
We will drop the `Handle` immediately, before we start
`pgb_writer.flush()`ing the responses.
Thereby letting go of our hold on the `GateGuard`, allowing the timeline
shutdown to complete while the page_service handler remains intact.
# Code Changes
* Reproducer & Regression Test
* Developed and proven to reproduce the issue in
https://github.com/neondatabase/neon/pull/10399
* Add a `Test` message to the pagestream protocol (`cfg(feature =
"testing")`).
* Drive-by minimal improvement to the parsing code, we now have a
`PagestreamFeMessageTag`.
* Refactor `pageserver/client` to allow sending and receiving
`page_service` requests independently.
* Add a Rust helper binary to produce situation (4) from above
* Rationale: (4) and (5) are the same bug class, we're holding a gate
open while `flush()`ing.
* Add a Python regression test that uses the helper binary to
demonstrate the problem.
* Fix
* Introduce and use `WeakHandle` as explained earlier.
* Replace the `shut_down` atomic with two enum states for `HandleInner`,
wrapped in a `Mutex`.
* To make `WeakHandle::upgrade()` and `Handle::downgrade()`
cache-efficient:
* Wrap the `Types::Timeline` in an `Arc`
* Wrap the `GateGuard` in an `Arc`
* The separate `Arc`s enable uncontended cloning of the timeline
reference in `upgrade()` and `downgrade()`.
If instead we were `Arc<Timeline>::clone`, different connection handlers
would be hitting the same cache line on every upgrade()/downgrade(),
causing contention.
* Please read the udpated module-level comment in `mod handle`
module-level comment for details.
# Testing & Performance
The reproducer test that failed before the changes now passes, and
obviously other tests are passing as well.
We'll do more testing in staging, where the issue happens every ~4h if
chaos migrations are enabled in storcon.
Existing perf testing will be sufficient, no perf degradation is
expected.
It's a few more alloctations due to the added Arc's, but, they're low
frequency.
# Appendix: Why Compute Sometimes Doesn't Read Responses
Remember, the whole problem surfaced because flush() was slow because
Compute was not reading responses. Why is that?
In short, the way the compute works, it only advances the page_service
protocol processing when it has an interest in data, i.e., when the
pagestore smgr is called to return pages.
Thus, if compute issues a bunch of requests as part of prefetch but then
it turns out it can service the query without reading those pages, it
may very well happen that these messages stay in the TCP until the next
smgr read happens, either in that session, or possibly in another
session.
If there’s too many unread responses in the TCP, the pageserver kernel
is going to backpressure into userspace, resulting in our stuck flush().
All of this stems from the way vanilla Postgres does prefetching and
"async IO":
it issues `fadvise()` to make the kernel do the IO in the background,
buffering results in the kernel page cache.
It then consumes the results through synchronous `read()` system calls,
which hopefully will be fast because of the `fadvise()`.
If it turns out that some / all of the prefetch results are not needed,
Postgres will not be issuing those `read()` system calls.
The kernel will eventually react to that by reusing page cache pages
that hold completed prefetched data.
Uncompleted prefetch requests may or may not be processed -- it's up to
the kernel.
In Neon, the smgr + Pageserver together take on the role of the kernel
in above paragraphs.
In the current implementation, all prefetches are sent as GetPage
requests to Pageserver.
The responses are only processed in the places where vanilla Postgres
would do the synchronous `read()` system call.
If we never get to that, the responses are queued inside the TCP
connection, which, once buffers run full, will backpressure into
Pageserver's sending code, i.e., the `pgb_writer.flush()` that was the
root cause of the problems we're fixing in this PR.
The extension now supports Postgres 17. The release also seems to be
binary compatible with the previous version.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
`test_storage_controller_node_deletion` sometimes failed because shards
were moving around during timeline creation, and neon_local isn't
tolerant of that. The movements were unexpected because the shards had
only just been created.
This was a regression from #9916Closes: #10383
## Summary of changes
- Make this test use multiple AZs -- this makes the storage controller's
scheduling reliably stable
Why this works: in #9916 , I made a simplifying assumption that we would
have multiple AZs to get nice stable scheduling -- it's much easier,
because each tenant has a well defined primary+secondary location when
they have an AZ preference and nodes have different AZs. Everything
still works if you don't have multiple AZs, but you just have this quirk
that sometimes the optimizer can disagree with initial scheduling, so
once in a while a shard moves after being created -- annoying for tests,
harmless IRL.
## Problem
All pageserver have the same application name which makes it hard to
distinguish them.
## Summary of changes
Include the node id in the application name sent to the safekeeper. This
should gives us
more visibility in logs. There's a few metrics that will increase in
cardinality by `pageserver_count`,
but that's fine.
## Problem
Node fills were limited to moving (total shards / node_count) shards. In
systems that aren't perfectly balanced already, that leads us to skip
migrating some of the shards that belong on this node, generating work
for the optimizer later to gradually move them back.
## Summary of changes
- Where a shard has a preferred AZ and is currently attached outside
this AZ, then always promote it during fill, irrespective of target fill
count
## Problem
We were comparing serialized configs from the database with serialized
configs from memory. If fields have been added/removed to TenantConfig,
this generates spurious consistency errors. This is fine in test
environments, but limits the usefulness of this debug API in the field.
Closes: https://github.com/neondatabase/neon/issues/10369
## Summary of changes
- Do a decode/encode cycle on the config before comparing it, so that it
will have exactly the expected fields.
## Problem
gc-compaction needs the partitioning data to decide the job split. This
refactor allows concurrent access/computing the partitioning.
## Summary of changes
Make `partitioning` an ArcSwap so that others can access the
partitioning while we compute it. Fully eliminate the `repartition is
called concurrently` warning when gc-compaction is going on.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Rename the safekeeper scheduling policy "disabled" to "pause".
A rename was requested in
https://github.com/neondatabase/neon/pull/10400#discussion_r1916259124,
as the "disabled" policy is meant to be analogous to the "pause" policy
for pageservers.
Also simplify the `SkSchedulingPolicyArg::from_str` function, relying on
the `from_str` implementation of `SkSchedulingPolicy`. Latter is used
for the database format as well, so it is quite stable. If we ever want
to change the UI, we'll need to duplicate the function again but this is
cheap.
## Problem
Threads spawned in `test_tenant_delete_races_timeline_creation` are not
joined before the test ends, and can generate
`PytestUnhandledThreadExceptionWarning` in other tests.
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-10419/12805365523/index.html#/testresult/53a72568acd04dbd
## Summary of changes
- Wrap threads in ThreadPoolExecutor which will join them before the
test ends
- Remove a spurious deletion call -- the background thread doing
deletion ought to succeed.
## Problem
When multiple changes are grouped in a merge group to be merged as part
of the merge queue, the changes might individually pass
`check-codestyle-rust` but not in their combined form.
## Summary of changes
- Move `check-codestyle-rust` into a reusable workflow that is called
from it's previous location in `build_and_test.yml`, and additionally
call it from `pre_merge_checks.yml`. The additional call does not run on
ARM, only x86, to ensure the merge queue continues being responsive.
- Trigger `pre_merge_checks.yml` on PRs that change any of the workflows
running in `pre_merge_checks.yml`, so that we get feedback on those
early an not only after trying to merge those changes.
This should fix the largest source of flakyness of
test_nbtree_pagesplit_cycleid.
## Problem
https://github.com/neondatabase/neon/issues/10390
## Summary of changes
By using a guaranteed-flushed LSN, we ensure that PS won't have to wait
forever.
(If it does wait forever, we know the issue can't be with Compute's WAL)
## Problem
As part of https://github.com/neondatabase/neon/issues/8614 we need to
pass options to START_WAL_PUSH.
## Summary of changes
Add two options. `allow_timeline_creation`, default true, disables
implicit timeline creation in the connection from compute. Eventually
such creation will be forbidden completely, but as we migrate to
configurations we need to support both: current mode and configurations
enabled where creation by compute is disabled.
`proto_version` specifies compute <-> sk protocol version. We have it
currently in the first greeting package also, but I plan to change tag
size from u64 to u8, which would make it hard to use. Command is more
appropriate place for it anyway.
This reduces pressure on the OS TCP read buffer by increasing the
moments we read data out of the receive buffer, and increasing the
number of bytes we can pull from that buffer when we do reads.
## Problem
A backend may not always consume its prefetch data quick enough
## Summary of changes
We add a new function `prefetch_pump_state` which pulls as many prefetch
requests from the OS TCP receive buffer as possible, but without
blocking.
This thus reduces pressure on OS-level TCP buffers, thus increasing
throughput by limiting throttling caused by full TCP buffers.
## Problem
part of https://github.com/neondatabase/neon/issues/9114
part of investigation of
https://github.com/neondatabase/neon/issues/10049
## Summary of changes
* If `cfg!(test) or cfg!(feature = testing)`, then we will always try
generating an image to ensure the history is replayable, but not put the
image layer into the final layer results, therefore discovering wrong
key history before we hit a read error.
* I suspect it's easier to trigger some races if gc-compaction is
continuously run on a timeline, so I increased the frequency to twice
per 10 churns.
* Also, create branches in gc-compaction smoke tests to get more test
coverage.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Arpad Müller <arpad@neon.tech>
## Problem
`postgres` is system database at neon, so we need to do `pg_restore`
into `neondb` instead
https://github.com/neondatabase/cloud/issues/22100
## Summary of changes
Changed fast_import a little bit:
1. After succesfull connection creating `neondb` in postgres instance
2. Changed restore connstring to use new db
3. Added optional `source_connection_string`, which allows to skip
`s3_prefix` and just connect directly.
4. Added `-i` that stops process until sigterm
## TODO
- [x] test image in cplane e2e
- [ ] Change import job image back to latest after this merged (partial
revert of https://github.com/neondatabase/cloud/pull/22338)
## Problem
When a pageserver is receiving high rates of requests, we don't have a
good way to efficiently discover what the client's access pattern is.
Closes: https://github.com/neondatabase/neon/issues/10275
## Summary of changes
- Add
`/v1/tenant/x/timeline/y/page_trace?size_limit_bytes=...&time_limit_secs=...`
API, which returns a binary buffer.
- Add `pagectl page-trace` tool to decode and analyze the output.
---------
Co-authored-by: Erik Grinaker <erik@neon.tech>
Implementing the last missing endpoint of #9981, this adds support to
set the scheduling policy of an individual safekeeper, as specified in
the RFC. However, unlike in the RFC we call the endpoint
`scheduling_policy` not `status`
Closes#9981.
As for why not use the upsert endpoint for this: we want to have the
safekeeper upsert endpoint be used for testing and for deploying new
safekeepers, but not for changes of the scheduling policy. We don't want
to change any of the other fields when marking a safekeeper as
decommissioned for example, so we'd have to first fetch them only to
then specify them again. Of course one can also design an endpoint where
one can omit any field and it doesn't get modified, but it's still not
great for observability to put everything into one big "change something
about this safekeeper" endpoint.
## Problem
Safekeepers currently decode and interpret WAL for each shard
separately.
This is wasteful in terms of CPU memory usage - we've seen this in
profiles.
## Summary of changes
Fan-out interpreted WAL to multiple shards.
The basic is that wal decoding and interpretation happens in a separate
tokio task and senders
attach to it. Senders only receive batches concerning their shard and
only past the Lsn they've last seen.
Fan-out is gated behind the `wal_reader_fanout` safekeeper flag
(disabled by default for now).
When fan-out is enabled, it might be desirable to control the absolute
delta between the
current position and a new shard's desired position (i.e. how far behind
or ahead a shard may be).
`max_delta_for_fanout` is a new optional safekeeper flag which dictates
whether to create a new
WAL reader or attach to the existing one. By default, this behaviour is
disabled. Let's consider enabling
it if we spot the need for it in the field.
## Testing
Tests passed [here](https://github.com/neondatabase/neon/pull/10301)
with wal reader fanout enabled
as of
34f6a71718.
Related: https://github.com/neondatabase/neon/issues/9337
Epic: https://github.com/neondatabase/neon/issues/9329
## Problem
https://github.com/neondatabase/neon/issues/9965
## Summary of changes
Add to safekeeper http endpoint to switch membership configuration. Also
add it to python client for tests, and add simple test itself.
## Problem
Successful `benchmarks` runs doesn't have enough visibility
Ref https://neondb.slack.com/archives/C069Z2199DL/p1736868055094539
## Summary of changes
- Report both successful and failed `benchmarks` to Slack
- Update `slackapi/slack-github-action` action
## Problem
Currently, we call `InterpretedWalRecord::from_bytes_filtered`
from each shard. To serve multiple shards at the same time,
the API needs to allow for enquiring about multiple shards.
## Summary of changes
This commit tweaks it a pretty brute force way. Naively, we could
just generate the shard for a key, but pre and post split shards
may be subscribed at the same time, so doing it efficiently is more
complex.
## Problem
https://github.com/neondatabase/neon/issues/9965
## Summary of changes
Add safekeeper membership configuration struct itself and storing it in
the control file. In passing also add creation timestamp to the control
file (there were cases where I wanted it in the past).
Remove obsolete unused PersistedPeerInfo struct from control file (still
keep it control_file_upgrade.rs to have it in old upgrade code).
Remove the binary representation of cfile in the roundtrip test.
Updating it is annoying, and we still test the actual roundtrip.
Also add configuration to timeline creation http request, currently used
only in one python test. In passing, slightly change LSNs meaning in the
request: normally start_lsn is passed (the same as ancestor_start_lsn in
similar pageserver call), but we allow specifying higher commit_lsn for
manual intervention if needed. Also when given LSN initialize
term_history with it.
## Problem
https://github.com/neondatabase/neon/pull/8455 wasn't specific enough on
migration from current situation to enabling generations.
## Summary of changes
Describe the missing parts, including control plane pushing generation
to compute, which also defines whether generations are enabled -- non
zero value does it.
## Problem
For large deployments, the `control/v1/tenant` listing API can time out
transmitting a monolithic serialized response.
## Summary of changes
- Add `limit` and `start_after` parameters to listing API
- Update storcon_cli to use these parameters and limit requests to 1000
items at a time
## Problem
With upload queue reordering in #10218, we can easily get into a
situation where multiple index uploads are queued back to back, which
can't be parallelized. This will happen e.g. when multiple layer flushes
enqueue layer/index/layer/index/... and the layers skip the queue and
are uploaded in parallel.
These index uploads will incur serial S3 roundtrip latencies, and may
block later operations.
Touches #10096.
## Summary of changes
When multiple back-to-back index uploads are ready to upload, only
upload the most recent index and drop the rest.
## Problem
The upload queue can currently schedule an arbitrary number of tasks.
This can both spawn an unbounded number of Tokio tasks, and also
significantly slow down upload queue scheduling as it's quadratic in
number of operations.
Touches #10096.
## Summary of changes
Limit the number of inprogress tasks to the remote storage upload
concurrency. While this concurrency limit is shared across all tenants,
there's certainly no point in scheduling more than this -- we could even
consider setting the limit lower, but don't for now to avoid
artificially constraining tenants.
By setting PATH in the 'pg-build' layer, all the extension build layers
will inherit. No need to pass PG_CONFIG to all the various make
invocations either: once pg_config is in PATH, the Makefiles will pick
it up from there.
## Problem
The upload queue currently sees significant head-of-line blocking. For
example, index uploads act as upload barriers, and for every layer flush
we schedule a layer and index upload, which effectively serializes layer
uploads.
Resolves#10096.
## Summary of changes
Allow upload queue operations to bypass the queue if they don't conflict
with preceding operations, increasing parallelism.
NB: the upload queue currently schedules an explicit barrier after every
layer flush as well (see #8550). This must be removed to enable
parallelism. This will require a better mechanism for compaction
backpressure, see e.g. #8390 or #5415.
## Problem
Since https://github.com/neondatabase/neon/pull/9916, the preferred AZ
of a tenant is much more impactful, and we would like to make it more
visible in tooling.
## Summary of changes
- Include AZ in node describe API
- Include AZ info in node & tenant outputs in CLI
- Add metrics for per-node shard counts, labelled by AZ
- Add a CLI for setting preferred AZ on a tenant
- Extend AZ-setting API+CLI to handle None for clearing preferred AZ
## Problem
Before this PR, the pagestream throttle was applied weighted on a
per-batch basis.
This had several problems:
1. The throttle occurence counters were only bumped by `1` instead of
`batch_size`.
2. The throttle wait time aggregator metric only counted one wait time,
irrespective
of `batch_size`. That makes sense in some ways of looking at it but not
in others.
3. If the last request in the batch runs into the throttle, the other
requests in the
batch are also throttled, i.e., over-throttling happens (theoretical,
didn't measure
it in practice).
## Solution
It occured to me that we can simply push the throttling upwards into
`pagestream_read_message`.
This has the added benefit that in pipeline mode, the `executor` stage
will, if it is idle,
steal whatever requests already made it into the `spsc_fold` and execute
them; before this
change, that was not the case - the throttling happened in the
`executor` stage instead of
the `batcher` stage.
## Code Changes
There are two changes in this PR:
1. Lifting up the throttling into the `pagestream_read_message` method.
2. Move the throttling metrics out of the `Throttle` type into
`SmgrOpMetrics`.
Unlike the other smgr metrics, throttling is per-tenant, hence the Arc.
3. Refactor the `SmgrOpTimer` implementation to account for the new
observation states,
and simplify its design.
4. Drive-by-fix flush time metrics. It was using the same `now` in the
`observe_guard` every time.
The `SmgrOpTimer` is now a state machine.
Each observation point moves the state machine forward.
If a timer object is dropped early some "pair"-like metrics still
require an increment or observation.
That's done in the Drop implementation, by driving the state machine to
completion.
## Problem
Discovered during the relation dir refactor work.
If we do not create images as in this patch, we would get two set of
image layers:
```
0000...METADATA_KEYS
0000...REL_KEYS
```
They overlap at the same LSN and would cause data loss for relation
keys. This doesn't happen in prod because initial image layer generation
is never called, but better to be fixed to avoid future issues with the
reldir refactors.
## Summary of changes
* Consolidate create_image_layers call into a single one.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Because of https://github.com/Azure/azure-sdk-for-rust/issues/1739, our
identity token file was not being refreshed. This caused our uploads to
start failing when the storage token expired.
## Summary of changes
Drop and recreate the remote storage config every time we upload in
order to force reload the identity token file.
## Problem
See https://github.com/neondatabase/neon/issues/10167
Too small number of `max_connections` (2) can cause failures of
test_physical_replication_config_mismatch_too_many_known_xids test
## Summary of changes
Increase `max_connections` to 5
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
We want to do a more robust job of scheduling tenants into their home
AZ: https://github.com/neondatabase/neon/issues/8264.
Closes: https://github.com/neondatabase/neon/issues/8969
## Summary of changes
### Scope
This PR combines prioritizing AZ with a larger rework of how we do
optimisation. The rationale is that just bumping AZ in the order of
Score attributes is a very tiny change: the interesting part is lining
up all the optimisation logic to respect this properly, which means
rewriting it to use the same scores as the scheduler, rather than the
fragile hand-crafted logic that we had before. Separating these changes
out is possible, but would involve doing two rounds of test updates
instead of one.
### Scheduling optimisation
`TenantShard`'s `optimize_attachment` and `optimize_secondary` methods
now both use the scheduler to pick a new "favourite" location. Then
there is some refined logic for whether + how to migrate to it:
- To decide if a new location is sufficiently "better", we generate
scores using some projected ScheduleContexts that exclude the shard
under consideration, so that we avoid migrating from a node with
AffinityScore(2) to a node with AffinityScore(1), only to migrate back
later.
- Score types get a `for_optimization` method so that when we compare
scores, we will only do an optimisation if the scores differ by their
highest-ranking attributes, not just because one pageserver is lower in
utilization. Eventually we _will_ want a mode that does this, but doing
it here would make scheduling logic unstable and harder to test, and to
do this correctly one needs to know the size of the tenant that one is
migrating.
- When we find a new attached location that we would like to move to, we
will create a new secondary location there, even if we already had one
on some other node. This handles the case where we have a home AZ A, and
want to migrate the attachment between pageservers in that AZ while
retaining a secondary location in some other AZ as well.
- A unit test is added for
https://github.com/neondatabase/neon/issues/8969, which is implicitly
fixed by reworking optimisation to use the same scheduling scores as
scheduling.
## Problem
In preparation to https://github.com/neondatabase/neon/issues/9516. We
need to store rel size and directory data in the sparse keyspace, but it
does not support inheritance yet.
## Summary of changes
Add a new type of keyspace "sparse but inherited" into the system.
On the read path: we don't remove the key range when we descend into the
ancestor. The search will stop when (1) the full key range is covered by
image layers (which has already been implemented before), or (2) we
reach the end of the ancestor chain.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Generally ed25519 seems to be much preferred for cryptographic strength
to P256 nowadays, and it is NIST approved finally. We should use it
where we can as it's also faster than p256.
This PR makes the re-signed JWTs between local_proxy and pg_session_jwt
use ed25519.
This does introduce a new dependency on ed25519, but I do recall some
Neon Authorise customers asking for support for ed25519, so I am
justifying this dependency addition in the context that we can then
introduce support for customer ed25519 keys
sources:
* https://csrc.nist.gov/pubs/fips/186-5/final subsection 7 (EdDSA)
* https://datatracker.ietf.org/doc/html/rfc8037#section-3.1
## Problem
Currently, if we want to move a secondary there isn't a neat way to do
that: we just have migration API for the attached location, and it is
only clean to use that if you've manually created a secondary via
pageserver API in the place you're going to move it to.
Secondary migration API enables:
- Moving the secondary somewhere because we would like to later move the
attached location there.
- Move the secondary location because we just want to reclaim some disk
space from its current location.
## Summary of changes
- Add `/migrate_secondary` API
- Add `tenant-shard-migrate-secondary` CLI
- Add tests for above
## Problem
We would sometimes fail to retry compute notifications:
1. Try and send, set compute_notify_failure if we can't
2. On next reconcile, reconcile() fails for some other reason (e.g.
tried to talk to an offline node), and we fail the `result.is_ok() &&
must_notify` condition around the re-sending.
Closes: https://github.com/neondatabase/cloud/issues/22612
## Summary of changes
- Clarify the meaning of the reconcile result: it should be Ok(()) if
configuring attached location worked, even if secondary or detach
locations cannot be reached.
- Skip trying to talk to secondaries if they're offline
- Even if reconcile fails and we can't send the compute notification (we
can't send it because we're not sure if it's really attached), make sure
we save the `compute_notify_failure` flag so that subsequent reconciler
runs will try again
- Add a regression test for the above
Taking continuous profiles every 20 seconds is likely too expensive (in
dollar terms). Let's try 60-second profiles. We can now interrupt
running profiles via `?force=true`, so this should be fine.
For testing the proxy's websockets support.
I wrote this to test https://github.com/neondatabase/neon/issues/3822.
Unfortunately, that bug can *not* be reproduced with this tunnel. The
bug only appears when the client pipelines the first query with the
authentication messages. The tunnel doesn't do that.
---
Update (@conradludgate 2025-01-10):
We have since added some websocket tests, but they manually implemented
a very simplistic setup of the postgres protocol. Introducing the tunnel
would make more complex testing simpler in the future.
---------
Co-authored-by: Conrad Ludgate <conradludgate@gmail.com>
## Problem
Limitations found while using this to investigate
https://github.com/neondatabase/neon/issues/10234:
- If we hit a node consistency issue, we drop out and don't check shards
for consistency
- The messages printed after a shard consistency issue are huge, and
grafana appears to drop them.
## Summary of changes
- Defer node consistency errors until the end of the function, so that
we always proceed to check shards for consistency
- Print out smaller log lines that just point out the diffs between
expected and persistent state
With a new beta build of the rust compiler, it's good to check out the
new lints. Either to find false positives, or find flaws in our code.
Additionally, it helps reduce the effort required to update to 1.85 in 6
weeks.
## Problem
It's only possible to take one CPU profile at a time. With Grafana
continuous profiling, a (low-frequency) CPU profile will always be
running, making it hard to take an ad hoc CPU profile at the same time.
Resolves#10072.
## Summary of changes
Add a `force` parameter for `/profile/cpu` which will end and return an
already running CPU profile, starting a new one for the current caller.
## Problem
Currently, the heap profiling frequency is every 1 MB allocated. Taking
a profile stack trace takes about 1 µs, and allocating 1 MB takes about
15 µs, so the overhead is about 6.7% which is a bit high. This is a
fixed cost regardless of whether heap profiles are actually accessed.
## Summary of changes
Increase the heap profiling sample frequency from 1 MB to 2 MB, which
reduces the overhead to about 3.3%. This seems acceptable, considering
performance-sensitive code will avoid allocations as far as possible
anyway.
There used to be some pg version dependencies in these extensions, but
now that there isn't, follow the simpler pattern used in other
extensions. No change in the produced images.
We don't need or want the `active` column. Remove it. Vlad pointed out
that this is safe.
Thanks to the separation of the schemata in earlier PRs, this is easy.
follow-up of #10205
Part of https://github.com/neondatabase/neon/issues/9981
When we moved throttling up from Timeline::get into page_service,
we stopped being sensitive to `Timeline::cancel`, even though we're
holding a Handle and thus a guard on the `Timeline::gate` open.
This PR rectifies the situation.
Refs
- Found while investigating #10309 (hung detach because gate kept open),
but not expected to be the root cause of that issue because the
affected tenants are not being throttled according to their metrics.
## Problem
https://github.com/neondatabase/infra/pull/2725 updated the scrubber to
use a non-host
port endpoint for storcon. That breaks when unwrapping the port.
## Summary of changes
Support both `host:port` and `host` formats for the storcon api.
## Problem
These two tests came up in #9537 as doing multi-gigabyte I/O, and from
inspection of the tests it doesn't seem like they need that to fulfil
their purpose.
## Summary of changes
- In test_local_file_cache_unlink, run fewer background threads with a
smaller number of rows. These background threads AFAICT exist to make
sure some I/O is going on while we unlink the LFC directory, but 5
threads should be enough for "some".
- In test_lfc_resize, tweak the test to validate that the cache size is
larger than the final size before resizing it, so that we're sure we're
writing enough data to really be doing something. Then decrease the
pgbench scale.
## Problem
When poetry v2 (released Jan 5) is used it needs `packaging.metadata`
module, but we downgrade `packaging` to 23.0. `packaging==23.1`
introduced the metadata submodule.
## Summary of changes
Update `packaging` to 24.2.
## Problem
This test writes ~5GB of data. It is not suitable to run in parallel
with all the other small tests in test_runner/regress.
via #9537
## Summary of changes
- Move test_parallel_copy into the performance directory, so that it
does not run in parallel with other tests
## Problem
I noticed in https://github.com/neondatabase/neon/pull/9537 that tests
which work with compat snapshots were writing several hundred MB of
data, which isn't really necessary.
Also, the snapshots are large but don't have the proper variety of
storage format features, e.g. they could just have L0 deltas.
## Summary of changes
- Use smaller scale factor and runtime to generate less data
- Configure a small layer size and use force image layer generation so
that our output contains L1 deltas and image layers, and has a decent
number of entries in the layer map
## Problem
Unlike CPU profiles, the `/profile/heap` endpoint can't automatically
generate SVG flamegraphs. This requires the user to install and use
`pprof` tooling, which is unnecessary and annoying.
Resolves#10203.
## Summary of changes
Add `format=svg` for the `/profile/heap` route, and generate an SVG
flamegraph using the `inferno` crate, similarly to what `pprof-rs`
already does for CPU profiles.
# Problem
Before this PR, there were cases where send() in state
SenderWaitsForReceiverToConsume would never be woken up
by the receiver, because it never registered with `wake_sender`.
Example Scenario 1: we stop polling a send() future A that was waiting
for the receiver to consume. We drop A and create a new send() future B.
B would return Poll::Pending and never regsister a waker.
Example Scenario 2: a send() future A transitions from HasData
to SenderWaitsForReceiverToConsume. This registers the context X
with `wake_sender`. But before the Receiver consumes the data,
we poll A from a different context Y.
The state is still SenderWaitsForReceiverToConsume, but we wouldn't
register the new context with `wake_sender`.
When the Receiver comes around to consume and `wake_sender.notify()`s,
it wakes the old context X instead of Y.
# Fix
Register the waker in the case where we're polled in
state `SenderWaitsForReceiverToConsume`.
# Relation to #10309
I found this bug while investigating #10309.
There was never proof that this bug here is the root cause for #10309.
In the meantime we found a more probably hypothesis
for the root cause than what is being fixed here.
Regardless, let's walk through my thought process about
how it might have been relevant:
There (in page_service), Scenario 1 does not apply because
we poll the send() future to completion.
Scenario 2 (`tokio::join!`) also does not apply with the
current `tokio::join!()` impl, because it will just poll each
future every time, each with the same context.
Although if we ever used something like a FuturesUnordered anywhere,
that will be using a different context, so, in that case,
the bug might materialize.
Regarding tokio & spurious poll in general:
@conradludgate is not aware of any spurious wakeup cases in current
tokio,
but within a `tokio::join!()`, any wake meant for one future will poll
all
the futures, so that can appear as a spurious wake up to the N-1 futures
of the `tokio::join!()`.
## Problem
We were incorrectly constructing the ComputeUserInfo, used for
cancellation checks, based on the return parameters from postgres. This
didn't contain the correct info.
## Summary of changes
Propagate down the existing ComputeUserInfo.
## Problem
When the proxy receives a `Notification` with an unknown topic it's
supposed to use the `UnknownTopic` unit variant. Unfortunately, in
adjacently tagged enums serde will not simply ignore the configured
content if found and try to deserialize a map/object instead.
## Summary of changes
* Use a custom deserialize function to ignore variant content.
* Add a little unit test covering both cases.
## Problem
Auto-offloading as requested by the compaction task is racy with
unarchival, in that the compaction task might attempt to offload an
unarchived timeline. By that point it will already have set the timeline
to the `Stopping` state however, which makes it unusable for any
purpose. For example:
1. compaction task decides to offload timeline
2. timeline gets unarchived
3. `offload_timeline` gets called by compaction task
* sets timeline's state to `Stopping`
* realizes that the timeline can't be unarchived, errors out
6. endpoint can't be started as the timeline is `Stopping` and thus
'can't be found'.
A future iteration of the compaction task can't "heal" this state either
as the timeline will still not be archived, same goes for other
automatic stuff. The only way to heal this is a tenant detach+attach, or
alternatively a pageserver restart.
Furthermore, the compaction task is especially amenable for such races
as it first stores `can_offload` into a variable, figures out whether
compaction is needed (which takes some time), and only then does it
attempt an offload operation: the time difference between "check" and
"use" is non-trivially small.
To make it even worse, we start the compaction task right after attach
of a tenant, and it is a common pattern by pageserver users to attach a
tenant to then immediately unarchive a timeline, so that an endpoint can
be started.
## Solutions not adopted
The simplest solution is to move the `can_offload` check to right before
attempting of the offload. But this is not a good solution, as no lock
is held between that check and timeline shutdown. So races would still
be possible, just become less likely.
I explored using the timeline state for this, as in adding an additional
enum variant. But `Timeline::set_state` is racy (#10297).
## Adopted solution
We use the lock on the timeline's upload queue as an arbiter: either
unarchival gets to it first and sours the state for auto-offloading, or
auto-offloading shuts it down, which stops any parallel unarchival in
its tracks. The key part is not releasing the upload queue's lock
between the check whether the timeline is archived or not, and shutting
it down (the actual implementation only sets `shutting_down` but it has
the same effect on `initialized_mut()` as a full shutdown). The rest of
the patch is stuff that follows from this.
We also move the part where we set the state to `Stopping` to after that
arbiter has decided the fate of the timeline. For deletions, we do keep
it inside `DeleteTimelineFlow::prepare` however, so that it is called
with all of the the timelines locks held that the function allocates
(timelines lock most importantly). This is only a precautionary measure
however, as I didn't want to analyze deletion related code for possible
races.
## Future changes
It might make sense to move `can_offload` to right before the offload
attempt. Maybe some other properties might have changed as well.
Although this will not be perfect either as no lock is held. I want to
keep it out of this change to emphasize that this move wasn't the main
reason we are race free now.
Fixes#10220
This is a refactor to create better abstractions related to our
management server. It cleans up the code, and prepares everything for
authorized communication to and from the control plane.
Signed-off-by: Tristan Partin <tristan@neon.tech>
We keep the practice of keeping the compiler up to date, pointing to the
latest release. This is done by many other projects in the Rust
ecosystem as well.
[Release notes](https://releases.rs/docs/1.84.0/).
Prior update was in #9926.
## Problem
In Postgres, one cannot drop a role if it has any dependent objects in
the DB. In `compute_ctl`, we automatically reassign all dependent
objects in every DB to the corresponding DB owner. Yet, it seems that it
doesn't help with some implicit permissions. The issue is reproduced by
installing a `postgis` extension because it creates some views and
tables in the public schema.
## Summary of changes
Added a repro test without using a `postgis`: i) create a role via
`compute_ctl` (with `neon_superuser` grant); ii) create a test role, a
table in schema public, and grant permissions via the role in
`neon_superuser`.
To fix the issue, I added a new `compute_ctl` code that removes such
dangling permissions before dropping the role. It's done in the least
invasive way, i.e., only touches the schema public, because i) that's
the problem we had with PostGIS; ii) it creates a smaller chance of
messing anything up and getting a stuck operation again, just for a
different reason.
Properly, any API-based catalog operations should fail gracefully and
provide an actionable error and status code to the control plane,
allowing the latter to unwind the operation and propagate an error
message and hint to the user. In this sense, it's aligned with another
feature request https://github.com/neondatabase/cloud/issues/21611Resolveneondatabase/cloud#13582
## Problem
Initially we defaulted this to zero to reduce risk. We have now been
using pooling in staging for some time without issues, so let's make it
the default for anyone using this software without setting the config
explicitly.
Closes: https://github.com/neondatabase/cloud/issues/20971
## Summary of changes
- Set Azure blob storage connection pool size to 8 by default
## Problem
Occasionally we see an unexpected error like:
```
ERROR spawn_heartbeat_driver: Failed to update node state 1 after heartbeat round: Shutting down\n')
Hint: use scripts/check_allowed_errors.sh to test any new allowed_error you add
```
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-10324/12690404952/index.html#/testresult/63406a0687bf6eca
## Summary of changes
- Explicitly handle ApiError::ShuttingDown as a no-op when mutating node
status
## Problem
If for some reasons we already garbage-collected the data under an LSN
but the caller uses a past LSN for the find_time_cutoff function, now we
will report a missing key error and GC will never proceed.
Note that missing key error can also happen if the key is really missing
(i.e., during the past offload incidents)
## Summary of changes
Make sure GC proceeds by bumping the LSN. When time_cutoff=None, we will
not increase the time_cutoff (it will be set to latest_gc_cutoff). If we
really need to bump the GC LSN for maintenance purpose, we need a
separate API to do that.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We have several serious data corruption incidents caused by mismatch of
get-age requests:
https://neondb.slack.com/archives/C07FJS4QF7V/p1723032720164359
We hope that the problem is fixed now. But it is better to prevent such
kind of problems in future.
Part of https://github.com/neondatabase/cloud/issues/16472
## Summary of changes
This PR introduce new V3 version of compute<->pageserver protocol,
adding tag to getpage response.
So now compute is able to check if it really gets response to the
requested page.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
## Problem
The filtered record metric doesn't make sense for interpreted ingest.
## Summary of changes
While of dubious utility in the first place, this patch replaces them
with records received and records observed metrics for interpreted
ingest:
* received records cause the pageserver to do _something_: write a key,
value pair to storage, update some metadata or flush pending
modifications
* observed records are a shard 0 concept and contain only key metadata
used in tracking relation sizes (received records include observed
records)
## Problem
We want to define the algorithm for safekeeper membership change.
## Summary of changes
Add spec for it, several models and logs of checking them.
ref https://github.com/neondatabase/neon/issues/8699
## Problem
This was causing storage controller to still use neon-built libpq
instead of vanilla libpq.
Since https://github.com/neondatabase/neon/pull/10269 we have a vanilla
postgres in the system path -- anything that wants a postgres library
will use that.
## Summary of changes
- Remove LD_LIBRARY_PATH assignment in Dockerfile
This PR removes the direct dependency of the IP allowlist from
CancelClosure, allowing for more scalable and flexible IP restrictions
and enabling the future use of Redis-based CancelMap storage.
Changes:
- Introduce a new BackendAuth async trait that retrieves the IP
allowlist through existing authentication methods;
- Improve cancellation error handling by instrument() async
cancel_sesion() rather than dropping it.
- Set and store IP allowlist for SCRAM Proxy to consistently perform IP
allowance check
Relates to #9660
## Problem
Project gets stuck if database with subscriptions was deleted via API /
UI.
https://github.com/neondatabase/cloud/issues/18646
## Summary of changes
Before dropping the database, drop all the subscriptions in it.
Do not drop slot on publisher, because we have no guarantee that the
slot still exists or that the publisher is reachable.
Add `DropSubscriptionsForDeletedDatabases` phase to run these operations
in all databases, we're about to delete.
Ignore the error if the database does not exist.
## Problem
Typical deployments of neon have some tenants that stay in use
continuously, and a background churning population of tenants that are
created and then fall idle, and are configured to Detached state.
Currently, this churn of short lived tenants results in an
ever-increasing memory footprint.
Closes: https://github.com/neondatabase/neon/issues/9712
## Summary of changes
- At startup, filter to only load shards that don't have Detached policy
- In process_result, check if a tenant's shards are all Detached and
observed=={}, and if so drop them from memory
- In tenant_location_conf and other tenant mutators, load the tenants'
shards on-demand if they are not present
## Problem
The observed state removal may race with the inline updates of the
observed state done from `Service::node_activate_reconcile`.
This was intended to work as follows:
1. Detaches while the node is unavailable remove the entry from the
observed state.
2. `Service::node_activate_reconcile` diffs the locations returned
by the pageserver with the observed state and detaches in-line
when required.
## Summary of changes
This PR removes step (1) and lets background reconciliations
deal with the mismatch between the intent and observed state.
A follow up will attempt to remove `Service::node_activate_reconcile`
altogether.
Closes https://github.com/neondatabase/neon/issues/10253
## Problem
Consider the pageserver is doing the following sequence of operations:
* upload X files
* update index_part to add X and remove Y
* delete Y files
When storage scrubber obtains the initial timeline snapshot before
"update index_part" (that is the old version that contains Y but not X),
and then obtains the index_part file after it gets updated, it will
report all Y files are missing.
## Summary of changes
Do not report layer file missing if index_part listed and downloaded are
not the same (i.e. different last_modified times)
Signed-off-by: Alex Chi Z <chi@neon.tech>
Closes: https://github.com/neondatabase/cloud/issues/17784
## Problem
Currently, we run the whole CI pipeline for any changes. It's slow and
expensive.
## Suggestion
Starting with MacOs builds:
- check what files were changed
- rebuild only needed parts
- reuse results from previous builds when available
- run builds in parallel when possible
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
## Problem
We currently parse Notification twice even in the happy path.
## Summary of changes
Use `#[serde(other)]` to catch unknown topics and defer the second
parsing.
## Problem
`promote-images` was split into `promote-images-dev` and
`promote-images-prod` in
https://github.com/neondatabase/neon/pull/10267.
`dev` credentials were loaded in `promote-images-dev` and `prod`
credentials were loaded in `promote-images-prod`, but
`promote-images-prod` needs `dev` credentials as well to access the
`dev` images to replicate them from `dev` to `prod`.
## Summary of changes
Load `dev` credentials in `promote-images-prod` as well.
Apparently, we failed to do this bookkeeping in quite a few places...
## Problem
Fixes https://github.com/neondatabase/cloud/issues/22364
## Summary of changes
Add accounting of dropped requests. Note that this includes prefetches
dropped due to things like "PS connection dropped unexpectedly" or
"prefetch queue is already full", but *not* (yet?) "dropped due to
backend shutdown".
## Problem
`trigger-e2e-tests` waits half an hour before starting to run. Nearly
half of that time can be saved by promoting images before tests on them
are complete, so the e2e tests can run in parallel.
On `main` and `release{,-proxy,-compute}`, `promote-images` updates
`latest` and pushes things to prod ecr, so we want to run
`promote-images` only after `test-images` is done, but on other
branches, there is no harm in promoting images that aren't tested yet.
## Summary of changes
To promote images into dev container registries sooner, `promote-images`
is split into `promote-images-dev` and `promote-images-prod`. The former
pushes to dev container registries, the latter to prod ones. The latter
also waits for `test-images`, while the former doesn't. This allows to
run `trigger-e2e-tests` sooner.
Using `min(0, ...)` causes us to fail to wait in most situations, so a
lack of data would be a hot wait loop, which is bad.
## Problem
We noticed high CPU usage in some situations
## Problem
On macOS:
```
error: unused variable: `disable_lfc_resizing`
--> compute_tools/src/bin/compute_ctl.rs:431:9
|
431 | disable_lfc_resizing,
| ^^^^^^^^^^^^^^^^^^^^ help: try ignoring the field: `disable_lfc_resizing: _`
|
= note: `-D unused-variables` implied by `-D warnings`
= help: to override `-D warnings` add `#[allow(unused_variables)]`
```
## Summary of changes
- Initialise `disable_lfc_resizing` only on Linux (because it's used on
Linux only in further bloc)
## Problem
It's impossible to run regression tests with Python 3.13 as some
dependencies don't support it (some of them are outdated, and `jsonnet`
doesn't support it at all yet)
## Summary of changes
- Update dependencies for Python 3.13
- Install `jsonnet` only on Python < 3.13 and skip relevant tests on
Python 3.13
Closes#10237
## Problem
close https://github.com/neondatabase/neon/issues/10192
## Summary of changes
* `find_gc_time_cutoff` takes `now` parameter so that all branches
compute the cutoff based on the same start time, avoiding races.
* gc-compaction uses a single `get_gc_compaction_watermark` function to
get the safe LSN to compact.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
Frame pointers are typically disabled by default (depending on CPU
architecture), to improve performance. This frees up a CPU register, and
avoids a couple of instructions per function call. However, it makes
stack unwinding much more inefficient, since it has to use DWARF debug
information instead, and gives worse results with e.g. `perf` and eBPF
profiles. The `backtrace` implementation of `libunwind` is also
suspected to cause seg faults.
The performance benefit of frame pointer omission doesn't appear to
matter that much on modern 64-bit CPU architectures (which have plenty
of registers and optimized instruction execution), and benchmarks did
not show measurable overhead.
The Rust standard library and jemalloc already enable frame pointers by
default.
For more information, see
https://www.brendangregg.com/blog/2024-03-17/the-return-of-the-frame-pointers.html.
Resolves#10224.
Touches #10225.
## Summary of changes
Enable frame pointers in all builds, and use frame pointers for pprof-rs
stack sampling.
## Problem
Before the holidays, and just before our code freeze, a change to cplane
was made that started publishing the topics from #10197. This triggered
our alerts and put us in a sticky situation as it was not an error, and
we didn't want to silence the alert for the entire holidays, and we
didn't want to release proxy 2 days in a row if it was not essential.
We fixed it eventually by rewriting the alert based on logs, but this is
not a good solution.
## Summary of changes
Introduces an intermediate parsing step to check the topic name first,
to allow us to ignore parsing errors for any topics we do not know
about.
## Problem
We are chasing down segfaults in the storage controller
https://github.com/neondatabase/cloud/issues/21010
This is for use by the storage controller, which links dynamically with
`libpq`. We currently use the neon-built libpq, but this may be unsafe
for use from multi-threaded programs like the controller, as it uses a
statically linked openssl
Precursor to https://github.com/neondatabase/neon/pull/10258
## Summary of changes
- Include `postgresql-15` in container builds.
The reason for using version 15 is simply because that is what's
available in Debian 12 without adding any extra repositories, and we
don't have any special need for latest version in our libpq usage.
## Problem
It's not legal to modify layers that are referenced by the current layer
index. Assert this in the upload queue, as preparation for upload queue
reordering.
Touches #10096.
## Summary of changes
Add a debug assertion that the upload queue does not modify layers
referenced by the current index.
I could be convinced that this should be a plain assertion, but will be
conservative for now.
## Problem
Since enabling continuous profiling in staging, we've seen frequent seg
faults. This is suspected to be because jemalloc and pprof-rs take a
stack trace at the same time, and the handlers aren't signal safe.
jemalloc does this probabilistically on every allocation, regardless of
whether someone is taking a heap profile, which means that any CPU
profile has a chance to cause a seg fault.
Touches #10225.
## Summary of changes
For now, just disable heap profiles -- CPU profiles are more important,
and we need to be able to take them without risking a crash.
There is a race condition between `Tenant::shutdown`'s `defuse_for_drop`
loop and `offload_timeline`, where timeline offloading can insert into a
tenant that is in the process of shutting down, in fact so far
progressed that the `defuse_for_drop` has already been called.
This prevents warn log lines of the form:
```
offloaded timeline <hash> was dropped without having cleaned it up at the ancestor
```
The solution piggybacks on the `offloaded_timelines` lock: both the
defuse loop and the offloaded timeline insertion need to acquire the
lock, and we know that the defuse loop only runs after the tenant has
set its `TenantState` to `Stopping`.
So if we hold the `offloaded_timelines` lock, and know that the
`TenantState` is not `Stopping`, then we know that the defuse loop has
not ran yet, and holding the lock ensures that it doesn't start running
while we are inserting the offloaded timeline.
Fixes#10070
## Problem
When we do a timeline CRUD operation, we check that the shards we need
to mutate are currently attached to a pageserver, by reading
`generation` and `generation_pageserver` from the database.
If any don't appear to be attached, we respond with a a 503 and "One or
more shards in tenant is not yet attached".
This is happening more often than expected, and it's not obvious with
current logging what's going on: specifically which shard has a problem,
and exactly what we're seeing in these persistent generation columns.
(Aside: it's possible that we broke something with the change in #10011
which clears generation_pageserver when we detach a shard, although if
so the mechanism isn't trivial: what should happen is that if we stamp
on generation_pageserver if a reconciler is running, then it shouldn't
matter because we're about to
## Summary of changes
- When we are in Attached mode but find that
generation_pageserver/generation are unset, output details while looping
over shards.
## Problem
We see periodic failures in `test_scrubber_physical_gc_ancestors`, where
the logs show that the pageserver is creating image layers that should
cause child shards to no longer reference their parents' layers, but
then the scrubber runs and doesn't find any unreferenced layers.[
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-10256/12582034135/index.html#/testresult/78ea06dea6ba8dd3
From inspecting the code & test, it seems like this could be as simple
as the test failing to wait for uploads before running the scrubber. It
had a 2 second delay built in to satisfy the scrubbers time threshold
checks, which on a lightly loaded machine would also have been easily
enough for uploads to complete, but our test machines are more heavily
loaded all the time.
## Summary of changes
- Wait for uploads to complete after generating images layers in
test_scrubber_physical_gc_ancestors, so that the scrubber should
reliably see the post-compaction metadata.
## Problem
Versions of `diesel` and `pq-sys` were somewhat stale. I was checking on
libpq->openssl versions while investigating a segfault via
https://github.com/neondatabase/cloud/issues/21010. I don't think these
rust bindings are likely to be the source of issues, but we might as
well freshen them as a precaution.
## Summary of changes
- Update diesel to 2.2.6
- Update pq-sys to 0.6.3
There was no value in saving them off to temporary variables.
Signed-off-by: Tristan Partin <tristan@neon.tech>
Signed-off-by: Tristan Partin <tristan@neon.tech>
ref neondatabase/cloud#21731
## Problem
When we manually override the LFC size for particular computes,
autoscaling will typically undo that because vm-monitor will resize LFC
itself.
So, we'd like a way to make vm-monitor not set LFC size — this actually
already exists, if we just don't give vm-monitor a postgres connection
string.
## Summary of changes
Add a new field to the compute spec, `disable_lfc_resizing`. When set to
`true`, we pass in `None` for its postgres connection string. That
matches the configuration tested in `neondatabase/autoscaling` CI.
## Problem
Building local_proxy and compute_tools features the same dependency
tree, but as they are currently built in separate clean layers all that
progress is wasted. For our arm builds that's an extra 10 minutes.
## Summary of changes
Combines the compute_tools and local_proxy build layers.
## Problem
https://neondb.slack.com/archives/C085MBDUSS2/p1734604792755369
## Summary of changes
Recognize and ignore the 3 new broadcast messages:
- `/block_public_or_vpc_access_updated`
- `/allowed_vpc_endpoints_updated_for_org`
- `/allowed_vpc_endpoints_updated_for_projects`
## Problem
Running clippy with `cargo hack --feature-powerset` in CI isn't
particularly fast. This PR follows-up on
https://github.com/neondatabase/neon/pull/8912 to improve the speed of
our clippy runs.
Parallelism as suggested in
https://github.com/neondatabase/neon/issues/9901 was tested, but didn't
show consistent enough improvements to be worth it. It actually
increased the amount of work done, as there's less cache hits when
clippy runs are spread out over multiple target directories.
Additionally, parallelism makes it so caching needs to be thought about
more actively and copying around target directories to enable
parallelism eats the rest of the performance gains from parallel
execution.
After some discussion, the decision was to instead cut down on the
number of jobs that are running further. The easiest way to do this is
to not run clippy *without* default features. The list of default
features is empty for all crates, and I haven't found anything using
`cfg(feature = "default")` either, so this is likely not going to change
anything except speeding the runs up.
## Summary of changes
Reduce the amount of feature combinations tried by `cargo hack` (as
suggested in
https://github.com/neondatabase/neon/pull/8912#pullrequestreview-2286482368)
by never disabling default features.
## Alternatives
- We can split things out into different jobs which reduces the time
until everything is finished by running more things in parallel. This
does however decreases the amount of cache hits and increases the amount
of time spent on overhead tasks like repo cloning and restoring caches
by doing those multiple times instead of once.
- We could replace `cargo hack [...] clippy` with `cargo clippy [...];
cargo clippy --features testing`. I'm not 100% sure how this compares to
the change here in the PR, but it does seem to run a bit faster. That
likely means it's doing less work, but without understanding what
exactly we loose by that I'd rather not do that for now. I'd appreciate
input on this though.
Now that we construct the TLS client config for cancellation as well as
connect, it feels appropriate to construct the same config once and
re-use it elsewhere. It might also help should #7500 require any extra
setup, so we can easily add it to all the appropriate call sites.
In #10207 it was clear there was some confusion with the current
connection logic. To analyse the flow to make sure there was no poll
stalling, I ended up with the following refactor.
Notable changes:
1. Now all functions called `poll_xyz` and that have a `cx: &mut
Context` argument must return a `Poll<_>` type, and can only return
`Pending` iff an internal poll call also returned `Pending`
2. State management is handled entirely by `poll_messages`. There are
now only 2 states which makes it much easier to keep track of.
Each commit should be self-reviewable and should be simple to verify
that it keeps the same behaviour
## Problem
Currently default value of storage controller heartbeat interval is
100msec. It means that 10 times per second it establish connection to
PS. And it seems to be quite expensive.
At MacOS right now storage_controller consumes 70% CPU and trusts - 30%.
So together they completely utilize one core.
A lot of us has Macs. Let's save environment a little bit and do not
waste electricity and contribute to global warming.
By the way, on prod we have interval 10seconds
## Summary of changes
Increase heartbeat interval from 100msec to 1 second.
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
In https://github.com/neondatabase/neon/pull/9897 we temporarily
disabled the layer valid check because the current one only considers
the end result of all compaction algorithms, but partial gc-compaction
would temporarily produce an "invalid" layer map.
part of https://github.com/neondatabase/neon/issues/9114
## Summary of changes
Allow LSN splits to overlap in the slow path check. Currently, the valid
check is only used in storage scrubber (background job) and during
gc-compaction (without taking layer lock). Therefore, it's fine for such
checks to be a little bit inefficient but more accurate.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
Safekeeper may currently send a batch to the pageserver even if it
hasn't decoded a new record.
I think this is quite unlikely in the field, but worth adressing.
## Summary of changes
Don't send anything if we haven't decoded a full record. Once this
merges and releases, the `InterpretedWalRecords` struct can be updated
to remove the Option wrapper for `next_record_lsn`.
## Problem
The benchmarking utilities are also useful for testing. We want to write
tests in the safekeeper crate.
## Summary of changes
This commit lifts the utils to the safekeeper crate. They are compiled
if the benchmarking features is enabled or if in test mode.
## Problem
test_timeline_archival_chaos does timeline creation with failure
injection, and thereby sometimes leaves timelines in a part created
state. This was being reported as corruption by the scrubber on test
teardown, because it considered a layer without an index to be an
invalid state. This was incorrect: the scrubber should accept this
state, it occurs legitimately during timeline creation.
Closes: https://github.com/neondatabase/neon/issues/9988
## Summary of changes
- Report a timeline with layers but no index as Relic rather than
MissingIndexPart.
- We retain the MissingIndexPart variant for the case where an index
_was_ found in the listing, but was not found by a subsequent GET, i.e.
racing with deletion.
## Problem
`test_pgdata_import_smoke` writes two gigabytes of pages and then reads
them back serially. This is CPU bottlenecked and results in a long
runtime, and sensitivity to CPU load from other tests on the same
machine.
Closes: https://github.com/neondatabase/neon/issues/10071
## Summary of changes
- Use effective_io_concurrency=32 when doing sequential scans through
2GiB of pages in test_pgdata_import_smoke. This is a ~10x runtime
decrease in the parts of the test that do sequential scans.
- Also set `effective_io_concurrency=2` for tests, as I noticed while
debugging that we were doing all getpage requests serially, which is bad
for checking the stability of the batching code.
## Problem
We want to verify how much / if pgbench throughput and latency on Neon
suffers if the database contains many other relations, too.
## Summary of changes
Modify the benchmarking.yml pgbench-compare job to
- create an addiitional project at scale factor 10 GiB
- before running pgbench add n tables (initially 10k) to the database
- then compare the pgbench throughput and latency to the existing
pgbench-compare at 10 Gib scale factor
We use a realistic template for the n relations that is a partitioned
table with some realistic data types, indexes and constraints - similar
to a table that we use internally.
Example run:
https://github.com/neondatabase/neon/actions/runs/12377565956/job/34547386959
## Problem
s5cmd doesn't pick up the pod service account
```
2024/12/16 16:26:01 Ignoring, HTTP credential provider invalid endpoint host, "169.254.170.23", only loopback hosts are allowed. <nil>
ERROR "ls s3://neon-dev-bulk-import-us-east-2/import-pgdata/fast-import/v1/br-wandering-hall-w2xobawv": NoCredentialProviders: no valid providers in chain. Deprecated. For verbose messaging see aws.Config.CredentialsChainVerboseErrors
```
## Summary of changes
Switch to offical CLI.
## Testing
Tested the pre-merge image in staging, using `job_image` override in
project settings.
https://neondb.slack.com/archives/C033RQ5SPDH/p1734554944391949?thread_ts=1734368383.258759&cid=C033RQ5SPDH
## Future Work
Switch back to s5cmd once https://github.com/peak/s5cmd/pull/769 gets
merged.
## Refs
- fixes https://github.com/neondatabase/cloud/issues/21876
---------
Co-authored-by: Gleb Novikov <NanoBjorn@users.noreply.github.com>
## Problem
In https://github.com/neondatabase/neon/pull/8103 we changed the test
case to have more test coverage of gc_compaction. Now that we have
`test_gc_compaction_smoke`, we can revert this test case to serve its
original purpose and revert the parameter changes.
part of https://github.com/neondatabase/neon/issues/9114
## Summary of changes
* Revert pitr_interval from 60s to 10s.
* Assert the physical/logical size ratio in the benchmark.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
We cannot get the size of the compaction queue and access the info.
Part of #9114
## Summary of changes
* Add an API endpoint to get the compaction queue.
* gc_compaction test case now waits until the compaction finishes.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
`neon_local` has always been unsafe to run concurrently with itself: it
uses simple text files for persistent state, and concurrent runs will
step on each other.
In some test environments we intentionally handle this with mutexes in
python land, but it's fragile to try and always remember to do that.
## Summary of changes
- Add a `flock` based mutex around the `main` function of neon_local,
using the repo directory as the file to lock
- Clean up an Option<> around control_plane_api, this is a drive-by
change because it was one of the fields that had a weird effect when
previous concurrent stuff stamped on it.
## Problem
part of https://github.com/neondatabase/neon/issues/9114
In https://github.com/neondatabase/neon/pull/10127 we fixed the race,
but we didn't add the errors to the allowlist.
## Summary of changes
* Allow repartition errors in the gc-compaction smoke test.
I think it might be worth to refactor the code to allow multiple threads
getting a copy of repartition status (i.e., using Rcu) in the future.
Signed-off-by: Alex Chi Z <chi@neon.tech>
Add a `safekeepers` subcommand to `storcon_cli` that allows listing the
safekeepers.
```
$ curl -X POST --url http://localhost:1234/control/v1/safekeeper/42 --data \
'{"active":true, "id":42, "created_at":"2023-10-25T09:11:25Z", "updated_at":"2024-08-28T11:32:43Z","region_id":"neon_local","host":"localhost","port":5454,"http_port":0,"version":123,"availability_zone_id":"us-east-2b"}'
$ cargo run --bin storcon_cli -- --api http://localhost:1234 safekeepers
Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.38s
Running `target/debug/storcon_cli --api 'http://localhost:1234' safekeepers`
+----+---------+-----------+------+-----------+------------+
| Id | Version | Host | Port | Http Port | AZ Id |
+==========================================================+
| 42 | 123 | localhost | 5454 | 0 | us-east-2b |
+----+---------+-----------+------+-----------+------------+
```
Also:
* Don't return the raw `SafekeeperPersistence` struct that contains the
raw database presentation, but instead a new
`SafekeeperDescribeResponse` struct.
* The `SafekeeperPersistence` struct leaves out the `active` field on
purpose because we want to deprecate it and replace it with a
`scheduling_policy` one.
Part of https://github.com/neondatabase/neon/issues/9981
## Problem
The allure report finishes with the error `HttpError: Resource not
accessible by integration` while running the `pg_regress` test against a
cloud staging project due to a lack of permissions.
## Summary of changes
The permissions are added.
## Problem
It is unreliable for the control plane to infer the AZ for computes from
where the tenant is currently attached, because if a tenant happens to
be in a degraded state or a release is ongoing while a compute starts,
then the tenant's attached AZ can be a different one to where it will
run long-term, and the control plane doesn't check back later to restart
the compute.
This can land in parallel with
https://github.com/neondatabase/neon/pull/9947
## Summary of changes
- Thread through the preferred AZ into the compute hook code via the
reconciler
- Include the preferred AZ in the body of compute hook notifications
## Problem
Jemalloc heap profiles aren't symbolized. This is inconvenient, and
doesn't work with Grafana Cloud Profiles.
Resolves#9964.
## Summary of changes
Symbolize the heap profiles in-process, and strip unnecessary cruft.
This uses about 100 MB additional memory to cache the DWARF information,
but I believe this is already the case with CPU profiles, which use the
same library for symbolization. With cached DWARF information, the
symbolization CPU overhead is negligible.
Example profiles:
*
[pageserver.pb.gz](https://github.com/user-attachments/files/18141395/pageserver.pb.gz)
*
[safekeeper.pb.gz](https://github.com/user-attachments/files/18141396/safekeeper.pb.gz)
Don't build tests in h3 and rdkit: ~15 min speedup.
Use Ninja as cmake generator where possible: ~10 min speedup.
Clean apt cache for smaller images: around 250mb size loss for
intermediate layers
## Problem
It was reported as `gauge`, but it's actually a `counter`.
Also add `_total` suffix as that's the convention for counters.
The corresponding flux-fleet PR:
https://github.com/neondatabase/flux-fleet/pull/386
## Problem
The ABS SDK's default behavior is to do no connection pooling, i.e. open
and close a fresh connection for each request. Under high request rates,
this can result in an accumulation of TCP connections in TIME_WAIT or
CLOSE_WAIT state, and in extreme cases exhaustion of client ports.
Related: https://github.com/neondatabase/cloud/issues/20971
## Summary of changes
- Add a configurable `conn_pool_size` parameter for Azure storage,
defaulting to zero (current behavior)
- Construct a custom reqwest client using this connection pool size.
## Problem
It's impossible to run docker compose with compute v17 due to `pg_anon`
extension which is not supported under PG17.
## Summary of changes
The auto-loading of `pg_anon` is disabled by default
## Problem
To debug issues with TLS connections there's no easy way to decrypt
packets unless a client has special support for logging the keys.
## Summary of changes
Add TLS session keys logging to proxy via `SSLKEYLOGFILE` env var gated
by flag.
As the title says, I updated the lint rules to no longer allow unwrap or
unimplemented.
Three special cases:
* Tests are allowed to use them
* std::sync::Mutex lock().unwrap() is common because it's usually
correct to continue panicking on poison
* `tokio::spawn_blocking(...).await.unwrap()` is common because it will
only error if the blocking fn panics, so continuing the panic is also
correct
I've introduced two extension traits to help with these last two, that
are a bit more explicit so they don't need an expect message every time.
## Problem
We've had similar test in test_logical_replication, but then removed it
because it wasn't needed to trigger LR related bug. Restarting at WAL
page boundary is still a useful test, so add it separately back.
## Summary of changes
Add the test.
## Problem
We want to use safekeeper http client in storage controller and
neon_local.
## Summary of changes
Extract it to separate crate. No functional changes.
## Problem
While reviewing #10152 I found it tricky to actually determine whether
the connection used `allow_self_signed_compute` or not.
I've tried to remove this setting in the past:
* https://github.com/neondatabase/neon/pull/7884
* https://github.com/neondatabase/neon/pull/7437
* https://github.com/neondatabase/cloud/pull/13702
But each time it seems it is used by e2e tests
## Summary of changes
The `node_info.allow_self_signed_computes` is always initialised to
false, and then sometimes inherits the proxy config value. There's no
need this needs to be in the node_info, so removing it and propagating
it via `TcpMechansim` is simpler.
## Problem
Changes in #9786 were functionally complete but missed some edges that
made testing less robust than it should have been:
- `is_key_disposable` didn't consider SLRU dir keys disposable
- Timeline `init_empty` was always creating SLRU dir keys on all shards
The result was that when we had a bug
(https://github.com/neondatabase/neon/pull/10080), it wasn't apparent in
tests, because one would only encounter the issue if running on a
long-lived timeline with enough compaction to drop the initially created
empty SLRU dir keys, _and_ some CLog truncation going on.
Closes: https://github.com/neondatabase/cloud/issues/21516
## Summary of changes
- Update is_key_global and init_empty to handle SLRU dir keys properly
-- the only functional impact is that we avoid writing some spurious
keys in shards >0, but this makes testing much more robust.
- Make `test_clog_truncate` explicitly use a sharded tenant
The net result is that if one reverts #10080, then tests fail (i.e. this
PR is a reproducer for the issue)
## Problem
In #8550, we made the flush loop wait for uploads after every layer.
This was to avoid unbounded buildup of uploads, and to reduce compaction
debt. However, the approach has several problems:
* It prevents upload parallelism.
* It prevents flush and upload pipelining.
* It slows down ingestion even when there is no need to backpressure.
* It does not directly backpressure WAL ingestion (only via
`disk_consistent_lsn`), and will build up in-memory layers.
* It does not directly backpressure based on compaction debt and read
amplification.
An alternative solution to these problems is proposed in #8390.
In the meanwhile, we revert the change to reduce the impact on ingest
throughput. This does reintroduce some risk of unbounded
upload/compaction buildup. Until
https://github.com/neondatabase/neon/issues/8390, this can be addressed
in other ways:
* Use `max_replication_apply_lag` (aka `remote_consistent_lsn`), which
will more directly limit upload debt.
* Shard the tenant, which will spread the flush/upload work across more
Pageservers and move the bottleneck to Safekeeper.
Touches #10095.
## Summary of changes
Remove waiting on the upload queue in the flush loop.
## Problem
When entry was dropped and password wasn't set, new entry
had uninitialized memory in controlplane adapter
Resolves: https://github.com/neondatabase/cloud/issues/14914
## Summary of changes
Initialize password in all cases, add tests.
Minor formatting for less indentation
## Problem
`benchmarking` job fails because `aws-oicd-role-arn` input is not set
## Summary of changes:
- Set `aws-oicd-role-arn` for `benchmarking job
- Always require `aws-oicd-role-arn` to be set
- Rename `aws_oicd_role_arn` to `aws-oicd-role-arn` for consistency
## Problem
close https://github.com/neondatabase/neon/issues/10124
gc-compaction split_gc_jobs is holding the repartition lock for too long
time.
## Summary of changes
* Ensure split_gc_compaction_jobs drops the repartition lock once it
finishes cloning the structures.
* Update comments.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Improved comments will help others when they read the code, and the log
messages will help others understand why the logical replication monitor
works the way it does.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
LFC used_pages statistic is not updated in case of LFC resize (shrinking
`neon.file_cache_size_limit`)
## Summary of changes
Update `lfc_ctl->used_pages` in `lfc_change_limit_hook`
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
The test was failing with the scary but generic message `Remote storage
metadata corrupted`.
The underlying scrubber error is `Orphan layer detected: ...`.
The test kills pageserver at random points, hence it's expected that we
leak layers if we're killed in the window after layer upload but before
it's referenced from index part.
Refer to generation numbers RFC for details.
Refs:
- fixes https://github.com/neondatabase/neon/issues/9988
- root-cause analysis
https://github.com/neondatabase/neon/issues/9988#issuecomment-2520673167
## Problem
`test_prefetch` is flaky
(https://github.com/neondatabase/neon/issues/9961), but if it passes,
the run time is less than 30 seconds — we don't need an extended timeout
for it.
## Summary of changes
- Remove extended test timeout for `test_prefetch`
## Problem
We want to extract safekeeper http client to separate crate for use in
storage controller and neon_local. However, many types used in the API
are internal to safekeeper.
## Summary of changes
Move them to safekeeper_api crate. No functional changes.
ref https://github.com/neondatabase/neon/issues/9011
## Problem
When moving the comment on proxy-releases from the yaml doc into a
javascript code block, I missed converting the comment marker from `#`
to `//`.
## Summary of changes
Correctly convert comment marker.
## Problem
I've noticed that debug builds with LFC fail more frequently and for
some reason ,their failure do block merging (but it should not)
## Summary of changes
- Do not run Debug builds with LFC
## Problem
When we update our scheduler/optimization code to respect AZs properly
(https://github.com/neondatabase/neon/pull/9916), the choice of AZ
becomes a much higher-stakes decision. We will pretty much always run a
tenant in its preferred AZ, and that AZ is fixed for the lifetime of the
tenant (unless a human intervenes)
Eventually, when we do auto-balancing based on utilization, I anticipate
that part of that will be to automatically change the AZ of tenants if
our original scheduling decisions have caused imbalance, but as an
interim measure, we can at least avoid making this scheduling decision
based purely on which AZ contains the emptiest node.
This is a precursor to https://github.com/neondatabase/neon/pull/9947
## Summary of changes
- When creating a tenant, instead of scheduling a shard and then reading
its preferred AZ back, make the AZ decision first.
- Instead of choosing AZ based on which node is emptiest, use the median
utilization of nodes in each AZ to pick the AZ to use. This avoids bad
AZ decisions during periods when some node has very low utilization
(such as after replacing a dead node)
I considered also making the selection a weighted pseudo-random choice
based on utilization, but wanted to avoid destabilising tests with that
for now.
## Problem
Now notifications about failures in `pg_regress` tests run on the
staging cloud instance, reach the channel `on-call-staging-stream`,
while they should reach `on-call-qa-staging-stream`
## Summary of changes
The channel changed.
## Problem
CI currently uses static credentials in some places. These are less
secure and hard to maintain, so we are going to deprecate them and use
OIDC auth.
## Summary of changes
- ci(fix): Use OIDC auth to upload artifact on s3
- ci(fix): Use OIDC auth to login on ECR
## Problem
Now that https://github.com/neondatabase/cloud/issues/15245 is done, we
can remove the old code.
## Summary of changes
Removes support for the ManagementV2 API, in favour of the ProxyV1 API.
## Problem
To give Storage more time on preprod — create a release branch on Friday
## Summary of changes
- Automatically create Storage release PR on Friday instead of Monday
This adds an API to the storage controller to list safekeepers
registered to it.
This PR does a `diesel print-schema > storage_controller/src/schema.rs`
because of an inconsistency between up.sql and schema.rs, introduced by
[this](2c142f14f7)
commit, so there is some updates of `schema.rs` due to that. As a
followup to this, we should maybe think about running `diesel
print-schema` in CI.
Part of #9981
## Problem
`test_check_visibility_map` has been seen to time out in debug tests.
## Summary of changes
Bump the timeout to 10 minutes (test reports indicate 7 minutes is
sufficient).
We don't want to disable the test entirely in debug builds, to exercise
this with debug assertions enabled.
Resolves#10069.
This adds some validation of invariants that we want to uphold wrt the
tenant manifest and `index_part.json`:
* the data the manifest has about a timeline must match with the data in
`index_part.json`. It might actually change, e.g. when we do reparenting
during detach ancestor, but that requires the timeline to be
unoffloaded, i.e. removed from the manifest.
* any timeline mentioned in index part, must, if present, be archived.
If we unarchive, we first update the tenant manifest to unoffload, and
only then update index part. And one needs to archive before offloading.
* it is legal for timelines to be mentioned in the manifest but have no
`index_part`: this is a temporary state visible during deletion of the
timeline. if the pageserver crashed, an attach of the tenant will clean
the state up.
* it is also legal for offloaded timelines to have an
`ancestor_retain_lsn` of None while having an `ancestor_timeline_id`.
This is for the to-be-added flattening functionality: the plan is to set
former to None if we have flattened a timeline.
follow-up of #9942
part of #8088
## Problem
We saw the drain/fill operations not drain fast enough in ap-southeast.
## Summary of changes
These are some quick changes to speed it up:
* double reconcile concurrency - this is now half of the available
reconcile bandwidth
* reduce the waiter polling timeout - this way we can spawn new
reconciliations faster
## Problem
Cplane and storage controller tenant config changes are not additive.
Any change overrides all existing tenant configs. This would be fine if
both did client side patching, but that's not the case.
Once this merges, we must update cplane to use the PATCH endpoint.
## Summary of changes
### High Level
Allow for patching of tenant configuration with a `PATCH
/v1/tenant/config` endpoint.
It takes the same data as it's PUT counterpart. For example the payload
below will update `gc_period` and unset `compaction_period`. All other
fields are left in their original state.
```
{
"tenant_id": "1234",
"gc_period": "10s",
"compaction_period": null
}
```
### Low Level
* PS and storcon gain `PATCH /v1/tenant/config` endpoints. PS endpoint
is only used for cplane managed instances.
* `storcon_cli` is updated to have separate commands for
`set-tenant-config` and `patch-tenant-config`
Related https://github.com/neondatabase/cloud/issues/21043
add owned_by_superuser field to filter out system extensions.
While on it, also correct related code:
- fix the metric setting: use set() instead of inc() in a loop.
inc() is not idempotent and can lead to incorrect results
if the function called multiple times. Currently it is only called at
compute start, but this will change soon.
- fix the return type of the installed_extensions endpoint
to match the metric. Currently it is only used in the test.
## Problem
Linking walproposer library (e.g. `cargo t`) produces linker errors:
/home/myrrc/neon/pgxn/neon/walproposer_compat.c:169: undefined reference
to `pg_snprintf'
The library with these symbols (libpgcommon.a) is present
## Summary of changes
Changed order of libraries resolution for linker
## Problem
We added support for LFC for tests but are still using it only for the
PG17 release.
## Summary of changes
LFC is enabled for all PG versions. Errors in tests with LFC enabled now
block merging as usual. We keep tests with disabled LFC for PG17
release. Tests on debug builds with LFC enabled still don't affect
permission to merge.
## Problem
If the control plane cannot be reached for some reason, compute_ctl
panics
## Summary of changes
panic is removed in favour of returning an error.
Code is reformatted a bit for more flat control flow
Resolves: #5391
## Problem
We get slru truncation commands on non-zero shards.
Compaction will drop the slru dir keys and ingest will fail when
receiving such records.
https://github.com/neondatabase/neon/pull/10080 fixed it for clog, but
not for multixact.
## Summary of changes
Only truncate multixact slrus on shard zero. I audited the rest of the
ingest code and it looks
fine from this pov.
## Problem
With pipelining enabled, the time a request spends in the batcher stage
counts towards the smgr op latency.
If pipelining is disabled, that time is not accounted for.
In practice, this results in a jump in smgr getpage latencies in various
dashboards and degrades the internal SLO.
## Solution
In a similar vein to #10042 and with a similar rationale, this PR stops
counting the time spent in batcher stage towards smgr op latency.
The smgr op latency metric is reduced to the actual execution time.
Time spent in batcher stage is tracked in a separate histogram.
I expect to remove that histogram after batching rollout is complete,
but it will be helpful in the meantime to reason about the rollout.
## Problem
Protobuf doesn't support 128 bit integers, so we encode the keys as two
64 bit integers. Issue is that when we split the 128 bit compact key we
use signed 64 bit integers to represent the two halves. This may result
in a negative lower half when relnode is larger than `0x00800000`. When
we convert the lower half to an i128 we get a negative `CompactKey`.
## Summary of Changes
Use unsigned integers when encoding into Protobuf.
## Deployment
* Prod: We disabled the interpreted proto, so no compat concerns.
* Staging: Disable the interpreted proto, do one release, and then
release the fixed version.
We do this because a negative int32 will convert to a large uint32 value
and could give
a key in the actual pageserver space. In production we would around this
by adding new
fields to the proto and deprecating the old ones, but we can make our
lives easy here.
* Pre-prod: Same as staging
## Problem
When dev deployments are disabled (or fail), the tags for releases
aren't created. It makes more sense to have tag and release creation
before the deployment to prevent situations like
[this](https://github.com/neondatabase/neon/pull/9959).
It is not enough to move the tag creation before the deployment. If the
deployment fails, re-running the job isn't possible because the API call
to create the tag will fail.
## Summary of changes
- Tag/Release creation now happens before the deployment
- The two steps for tag and release have been merged into a bigger one
- There's new checks to ensure the that if the tags/releases already
exist as expected, things will continue just fine.
## Problem
In #9786 we stop storing SLRUs on non-zero shards.
However, there was one code path during ingest that still tries to
enumerate SLRU relations on all shards. This fails if it sees a tenant
who has never seen any write to an SLRU, or who has done such thorough
compaction+GC that it has dropped its SLRU directory key.
## Summary of changes
- Avoid trying to list SLRU relations on nonzero shards
Neon doesn't have seqscan detection of its own, so stop read_stream from
trying to utilize that readahead, and instead make it issue readahead of
its own.
## Problem
@knizhnik noticed that we didn't issue smgrprefetch[v] calls for
seqscans in PG17 due to the move to the read_stream API, which assumes
that the underlying IO facilities do seqscan detection for readahead.
That is a wrong assumption when Neon is involved, so let's remove the
code that applies that assumption.
## Summary of changes
Remove the cases where seqscans are detected and prefetch is disabled as
a consequence, and instead don't do that detection.
PG PR: https://github.com/neondatabase/postgres/pull/532
## Problem
resolve
https://github.com/neondatabase/neon/issues/9988#issuecomment-2528239437
## Summary of changes
* New verbose mode for storage scrubber scan metadata (pageserver) that
contains the error messages.
* Filter allowed_error list from the JSON output to determine the
healthy flag status.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
We have metrics for GetPage request latencies, but this is an extra
measure to capture requests that take way too long in the logs. The log
message is printed every 10 s, until the response is received:
```
PG:2024-12-09 16:02:07.715 GMT [1782845] LOG: [NEON_SMGR] [shard 0] no response received from pageserver for 10.000 s, still waiting (sent 10613 requests, received 10612 responses)
PG:2024-12-09 16:02:17.723 GMT [1782845] LOG: [NEON_SMGR] [shard 0] no response received from pageserver for 20.008 s, still waiting (sent 10613 requests, received 10612 responses)
PG:2024-12-09 16:02:19.719 GMT [1782845] LOG: [NEON_SMGR] [shard 0] received response from pageserver after 22.006 s
```
## Problem
close https://github.com/neondatabase/cloud/issues/19671
```
Timeline -----------------------------
^ last GC happened LSN
^ original retention period setting = 24hr
> refresh-gc-info updates the gc_info
^ planned cutoff (gc_info)
^ customer set retention to 48hr, and it's still within the last GC LSN
^1 ^2 we have two choices: (1) update the planned cutoff to
move backwards, or (2) keep the current one
```
In this patch, we decided to keep the current cutoff instead of moving
back the gc_info to avoid races. In the future, we could allow the
planned gc cutoff to go back once cplane sends a retention_history
tenant config update, but this requires a careful revisit of the code.
## Summary of changes
Ensure that GC cutoffs never go back if retention settings get changed.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
See https://github.com/neondatabase/neon/issues/9961
Current implementation of prefetch buffer resize doesn't correctly
handle in-flight requests
## Summary of changes
1. Fix index of entry we should wait for if new prefetch buffer size is
smaller than number of in-flight requests.
2. Correctly set flush position
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
Azure has a different per-request limit of 256 items for bulk deletion
compared to the number of 1000 on AWS. Therefore, we need to support
multiple values. Due to `GenericRemoteStorage`, we can't add an
associated constant, but it has to be a function.
The PR replaces the `MAX_KEYS_PER_DELETE` constant with a function of
the same name, implemented on both the `RemoteStorage` trait as well as
on `GenericRemoteStorage`.
The value serves as hint of how many objects to pass to the
`delete_objects` function.
Reading:
* https://learn.microsoft.com/en-us/rest/api/storageservices/blob-batch
* https://docs.aws.amazon.com/AmazonS3/latest/API/API_DeleteObjects.html
Part of #7931
Hello! I was interested in potentially making some contributions to Neon
and looking through the issue backlog I found
[8200](https://github.com/neondatabase/neon/issues/8200) which seemed
like a good first issue to attempt to tackle. I see it was assigned a
while ago so apologies if I'm stepping on any toes with this PR. I also
apologize for the size of this PR. I'm not sure if there is a simple way
to reduce it given the footprint of the components being changed.
## Problem
This PR is attempting to address part of the problem outlined in issue
[8200](https://github.com/neondatabase/neon/issues/8200). Namely to
remove global static usage of timeline state in favour of
`Arc<GlobalTimelines>` and to replace wasteful clones of
`SafeKeeperConf` with `Arc<SafeKeeperConf>`. I did not opt to tackle
`RemoteStorage` in this PR to minimize the amount of changes as this PR
is already quite large. I also did not opt to introduce an
`SafekeeperApp` wrapper struct to similarly minimize changes but I can
tackle either or both of these omissions in this PR if folks would like.
## Summary of changes
- Remove static usage of `GlobalTimelines` in favour of
`Arc<GlobalTimelines>`
- Wrap `SafeKeeperConf` in `Arc` to avoid wasteful clones of the
underlying struct
## Some additional thoughts
- We seem to currently store `SafeKeeperConf` in `GlobalTimelines` and
then expose it through a public`get_global_config` function which
requires locking. This seems needlessly wasteful and based on observed
usage we could remove this public accessor and force consumers to
acquire `SafeKeeperConf` through the new Arc reference.
## Problem
close https://github.com/neondatabase/neon/issues/10049, close
https://github.com/neondatabase/neon/issues/10030, close
https://github.com/neondatabase/neon/issues/8861
part of https://github.com/neondatabase/neon/issues/9114
The legacy gc process calls `get_latest_gc_cutoff`, which uses a Rcu
different than the gc_info struct. In the gc_compaction_smoke test case,
the "latest" cutoff could be lower than the gc_info struct, causing
gc-compaction to collect data that could be accessed by
`latest_gc_cutoff`. Technically speaking, there's nothing wrong with
gc-compaction using gc_info without considering latest_gc_cutoff,
because gc_info is the source of truth. But anyways, let's fix it.
## Summary of changes
* gc-compaction uses `latest_gc_cutoff` instead of gc_info to determine
the gc horizon.
* if a gc-compaction is scheduled via tenant compaction iteration, it
will take the gc_block lock to avoid racing with functionalities like
detach ancestor (if it's triggered via manual compaction API without
scheduling, then it won't take the lock)
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
Currently, we run the `pg_regress` tests only for PG16
However, PG17 is a part of Neon and should be tested as well
## Summary of changes
Modified the workflow and added a patch for PG17 enabling the
`pg_regress` tests.
The problem with leftovers was solved by using branches.
For a while already, we've been unable to update the Azure SDK crates
due to Azure adopting use of a non-tokio async runtime, see #7545.
The effort to upstream the fix got stalled, and I think it's better to
switch to a patched version of the SDK that is up to date.
Now we have a fork of the SDK under the neondatabase github org, to
which I have applied Conrad's rebased patches to:
https://github.com/neondatabase/azure-sdk-for-rust/tree/neon .
The existence of a fork will also help with shipping bulk delete support
before it's upstreamed (#7931).
Also, in related news, the Azure SDK has gotten a rift in development,
where the main branch pertains to a future, to-be-officially-blessed
release of the SDK, and the older versions, which we are currently
using, are on the `legacy` branch. Upstream doesn't really want patches
for the `legacy` branch any more, they want to focus on the `main`
efforts. However, even then, the `legacy` branch is still newer than
what we are having right now, so let's switch to `legacy` for now.
Depending on how long it takes, we can switch to the official version of
the SDK once it's released or switch to the upstream `main` branch if
there is changes we want before that.
As a nice side effect of this PR, we now use reqwest 0.12 everywhere,
dropping the dependency on version 0.11.
Fixes#7545
Result of running:
cargo update -p aws-types -p aws-sigv4 -p aws-credential-types -p
aws-smithy-types -p aws-smithy-async -p aws-sdk-kms -p aws-sdk-iam -p
aws-sdk-s3 -p aws-config
We want to keep the AWS SDK up to date as that way we benefit from new
developments and improvements.
## Problem
We saw a tenant get stuck when it had been put into Pause scheduling
mode to pin it to a pageserver, then it was left idle for a while and
the control plane tried to detach it.
Close: https://github.com/neondatabase/neon/issues/9957
## Summary of changes
- When changing policy to Detached or Secondary, set the scheduling
policy to Active.
- Add a test that exercises this
- When persisting tenant shards, set their `generation_pageserver` to
null if the placement policy is not Attached (this enables consistency
checks to work, and avoids leaving state in the DB that could be
confusing/misleading in future)
## Problem
In #9962 I changed the smgr metrics to include time spent on flush.
It isn't under our (=storage team's) control how long that flush takes
because the client can stop reading requests.
## Summary of changes
Stop the timer as soon as we've buffered up the response in the
`pgb_writer`.
Track flush time in a separate metric.
---------
Co-authored-by: Yuchen Liang <70461588+yliang412@users.noreply.github.com>
If the pageserver connection is lost while receiving the prefetch
request, the prefetch queue is cleared. The error message prints the
values from the prefetch slot, but because the slot was already cleared,
they're all zeros:
LOG: [NEON_SMGR] [shard 0] No response from reading prefetch entry 0:
0/0/0.0 block 0. This can be caused by a concurrent disconnect
To fix, make local copies of the values.
In the passing, also add a sanity check that if the receive() call
succeeds, the prefetch slot is still intact.
## Problem
part of https://github.com/neondatabase/neon/issues/9114, stacked PR
over #9809
The compaction scheduler now schedules partial compaction jobs.
## Summary of changes
* Add the compaction job splitter based on size.
* Schedule subcompactions using the compaction scheduler.
* Test subcompaction scheduler in the smoke regress test.
* Temporarily disable layer map checks
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
With the current metrics we can't identify which shards are ingesting
data at any given time.
## Summary of changes
Add a metric for the number of wal records received for processing by
each shard. This is per (tenant, timeline, shard).
## Problem
We didn't have a codeowner for `/compute`, so nobody was auto-assigned
for PRs like #9973
## Summary of changes
While on it:
1. Group codeowners into sections.
2. Remove control plane from the `/compute_tools` because it's primarily
the internal `compute_ctl` code.
3. Add control plane (and compute) to `/libs/compute_api` because that's
the shared public interface of the compute.
We've seen cases where stray keys end up on the wrong shard. This
shouldn't happen. Add debug assertions to prevent this. In release
builds, we should be lenient in order to handle changing key ownership
policies.
Touches #9914.
## Problem
There's no metrics for disk consistent LSN and remote LSN. This stuff is
useful when looking at ingest performance.
## Summary of changes
Two per timeline metrics are added: `pageserver_disk_consistent_lsn` and
`pageserver_projected_remote_consistent_lsn`. I went for the projected
remote lsn instead of the visible one
because that more closely matches remote storage write tput. Ideally we
would have both, but these metrics are expensive.
## Problem
I'm writing an ingest benchmark in #9812. To time S3 uploads, I need to
schedule a flush of the Pageserver's in-memory layer, but don't actually
want to wait around for it to complete (which will take a minute).
## Summary of changes
Add a parameter `wait_until_flush` (default `true`) for
`timeline/checkpoint` to control whether to wait for the flush to
complete.
## Problem
FSM pages are managed like regular relation pages, and owned by a single
shard. However, when truncating the FSM relation the last FSM page was
zeroed out on all shards. This is unnecessary and potentially confusing.
The superfluous keys will be removed during compactions, as they do not
belong on these shards.
Resolves#10027.
## Summary of changes
Only zero out the truncated FSM page on the owning shard.
## Problem
part of https://github.com/neondatabase/neon/issues/9114
gc-compaction can take a long time. This patch adds support for
scheduling a gc-compaction job. The compaction loop will first handle
L0->L1 compaction, and then gc compaction. The scheduled jobs are stored
in a non-persistent queue within the tenant structure.
This will be the building block for the partial compaction trigger -- if
the system determines that we need to do a gc compaction, it will
partition the keyspace and schedule several jobs. Each of these jobs
will run for a short amount of time (i.e, 1 min). L0 compaction will be
prioritized over gc compaction.
## Summary of changes
* Add compaction scheduler in tenant.
* Run scheduled compaction in integration tests.
* Change the manual compaction API to allow schedule a compaction
instead of immediately doing it.
* Add LSN upper bound as gc-compaction parameter. If we schedule partial
compactions, gc_cutoff might move across different runs. Therefore, we
need to pass a pre-determined gc_cutoff beforehand. (TODO: support LSN
lower bound so that we can compact arbitrary "rectangle" in the layer
map)
* Refactor the gc_compaction internal interface.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
We need a higher concurrency during reconfiguration in case of many DBs,
but the instance is already running and used by the client. We can
easily get out of `max_connections` limit, and the current code won't
handle that.
## Summary of changes
Default to 1, but also allow control plane to override this value for
specific projects. It's also recommended to bump
`superuser_reserved_connections` += `reconfigure_concurrency` for such
projects to ensure that we always have enough spare connections for
reconfiguration process to succeed.
Quick workaround for neondatabase/cloud#17846
## Problem
The node shard scan timeout of 1 second is a bit too aggressive, and
we've seen this cause test failures. The scans are performed in parallel
across nodes, and the entire operation has a 15 second timeout.
Resolves#9801.
## Summary of changes
Increase the timeout to 5 seconds. This is still enough to time out on a
network failure and retry successfully within 15 seconds.
Like #9931 but without rebasing upstream just yet, to try and minimise
the differences.
Removes all proxy-specific commits from the rust-postgres fork, now that
proxy no longer depends on them. Merging upstream changes to come later.
Closes#9387.
## Problem
`BufferedWriter` cannot proceed while the owned buffer is flushing to
disk. We want to implement double buffering so that the flush can happen
in the background. See #9387.
## Summary of changes
- Maintain two owned buffers in `BufferedWriter`.
- The writer is in charge of copying the data into owned, aligned
buffer, once full, submit it to the flush task.
- The flush background task is in charge of flushing the owned buffer to
disk, and returned the buffer to the writer for reuse.
- The writer and the flush background task communicate through a
bi-directional channel.
For in-memory layer, we also need to be able to read from the buffered
writer in `get_values_reconstruct_data`. To handle this case, we did the
following
- Use replace `VirtualFile::write_all` with `VirtualFile::write_all_at`,
and use `Arc` to share it between writer and background task.
- leverage `IoBufferMut::freeze` to get a cheaply clonable `IoBuffer`,
one clone will be submitted to the channel, the other clone will be
saved within the writer to serve reads. When we want to reuse the
buffer, we can invoke `IoBuffer::into_mut`, which gives us back the
mutable aligned buffer.
- InMemoryLayer reads is now aware of the maybe_flushed part of the
buffer.
**Caveat**
- We removed the owned version of write, because this interface does not
work well with buffer alignment. The result is that without direct IO
enabled,
[`download_object`](a439d57050/pageserver/src/tenant/remote_timeline_client/download.rs (L243))
does one more memcpy than before this PR due to the switch to use
`_borrowed` version of the write.
- "Bypass aligned part of write" could be implemented later to avoid
large amount of memcpy.
**Testing**
- use an oneshot channel based control mechanism to make flush behavior
deterministic in test.
- test reading from `EphemeralFile` when the last submitted buffer is
not flushed, in-progress, and done flushing to disk.
## Performance
We see performance improvement for small values, and regression on big
values, likely due to being CPU bound + disk write latency.
[Results](https://www.notion.so/neondatabase/Benchmarking-New-BufferedWriter-11-20-2024-143f189e0047805ba99acda89f984d51?pvs=4)
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
We have a scale test for the storage controller which also acts as a
good stress test for scheduling stability. However, it created nodes
with no AZs set.
## Summary of changes
- Bump node count to 6 and set AZs on them.
This is a precursor to other AZ-related PRs, to make sure any new code
that's landed is getting scale tested in an AZ-aware environment.
## Problem
We practice a manual release flow for the compute module. This will
allow automation of the compute release process.
## Summary of changes
The workflow was modified to make a compute release automatically on the
branch release-compute.
## Checklist before requesting a review
- [x] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
## Problem
Reqwest errors don't include details about the inner source error. This
means that we get opaque errors like:
```
receive body: error sending request for url (http://localhost:9898/v1/location_config)
```
Instead of the more helpful:
```
receive body: error sending request for url (http://localhost:9898/v1/location_config): operation timed out
```
Touches #9801.
## Summary of changes
Include the source error for `reqwest::Error` wherever it's displayed.
## Problem
When client specifies `application_name`, pgbouncer propagates it to the
Postgres. Yet, if client doesn't do it, we have hard time figuring out
who opens a lot of Postgres connections (including the `cloud_admin`
ones).
See this investigation as an example:
https://neondb.slack.com/archives/C0836R0RZ0D
## Summary of changes
I haven't found this documented, but it looks like pgbouncer accepts
standard Postgres connstring parameters in the connstring in the
`[databases]` section, so put the default `application_name=pgbouncer`
there. That way, we will always see who opens Postgres connections. I
did tests, and if client specifies a `application_name`, pgbouncer
overrides this default, so it only works if it's not specified or set to
blank `&application_name=` in the connection string.
This is the last place we could potentially open some Postgres
connections without `application_name`. Everything else should be either
of two:
1. Direct client connections without `application_name`, but these
should be strictly non-`cloud_admin` ones
2. Some ad-hoc internal connections, so if we see spikes of unidentified
`cloud_admin` connections, we will need to investigate it again.
Fixesneondatabase/cloud#20948
(stacked on #9990 and #9995)
Partially fixes#1287 with a custom option field to enable the fixed
behaviour. This allows us to gradually roll out the fix without silently
changing the observed behaviour for our customers.
related to https://github.com/neondatabase/cloud/issues/15284
## Problem
During deploys, we see a lot of 500 errors due to heapmap uploads for
inactive tenants. These should be 503s instead.
Resolves#9574.
## Summary of changes
Make the secondary tenant scheduler use `ApiError` rather than
`anyhow::Error`, to propagate the tenant error and convert it to an
appropriate status code.
## Problem
we tried different parallelism settings for ingest bench
## Summary of changes
the following settings seem optimal after merging
- SK side Wal filtering
- batched getpages
Settings:
- effective_io_concurrency 100
- concurrency limit 200 (different from Prod!)
- jobs 4, maintenance workers 7
- 10 GB chunk size
## Problem
```
2024-12-03T15:42:46.5978335Z + poetry run python /__w/neon/neon/scripts/ingest_perf_test_result.py --ingest /__w/neon/neon/test_runner/perf-report-local
2024-12-03T15:42:49.5325077Z Traceback (most recent call last):
2024-12-03T15:42:49.5325603Z File "/__w/neon/neon/scripts/ingest_perf_test_result.py", line 165, in <module>
2024-12-03T15:42:49.5326029Z main()
2024-12-03T15:42:49.5326316Z File "/__w/neon/neon/scripts/ingest_perf_test_result.py", line 155, in main
2024-12-03T15:42:49.5326739Z ingested = ingest_perf_test_result(cur, item, recorded_at_timestamp)
2024-12-03T15:42:49.5327488Z ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2024-12-03T15:42:49.5327914Z File "/__w/neon/neon/scripts/ingest_perf_test_result.py", line 99, in ingest_perf_test_result
2024-12-03T15:42:49.5328321Z psycopg2.extras.execute_values(
2024-12-03T15:42:49.5328940Z File "/github/home/.cache/pypoetry/virtualenvs/non-package-mode-_pxWMzVK-py3.11/lib/python3.11/site-packages/psycopg2/extras.py", line 1299, in execute_values
2024-12-03T15:42:49.5335618Z cur.execute(b''.join(parts))
2024-12-03T15:42:49.5335967Z psycopg2.errors.InvalidTextRepresentation: invalid input syntax for type numeric: "concurrent-futures"
2024-12-03T15:42:49.5336287Z LINE 57: 'concurrent-futures',
2024-12-03T15:42:49.5336462Z ^
```
## Summary of changes
- `test_page_service_batching`: save non-numeric params as `labels`
- Add a runtime check that `metric_value` is NUMERIC
Before this PR, some override callbacks used `.default()`, others
used `.setdefault()`.
As of this PR, all callbacks use `.setdefault()` which I think is least
prone to failure.
Aligning on a single way will set the right example for future tests
that need such customization.
The `test_pageserver_getpage_throttle.py` technically is a change in
behavior: before, it replaced the `tenant_config` field, now it just
configures the throttle. This is what I believe is intended anyway.
Support tenant manifests in the storage scrubber:
* list the manifests, order them by generation
* delete all manifests except for the two most recent generations
* for the latest manifest: try parsing it.
I've tested this patch by running the against a staging bucket and it
successfully deleted stuff (and avoided deleting the latest two
generations).
In follow-up work, we might want to also check some invariants of the
manifest, as mentioned in #8088.
Part of #9386
Part of #8088
---------
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
The Pageserver signal handler would only respond to a single signal and
initiate shutdown. Subsequent signals were ignored. This meant that a
`SIGQUIT` sent after a `SIGTERM` had no effect (e.g. in the case of a
slow or stalled shutdown). The `test_runner` uses this to force shutdown
if graceful shutdown is slow.
Touches #9740.
## Summary of changes
Keep responding to signals after the initial shutdown signal has been
received.
Arguably, the `test_runner` should also use `SIGKILL` rather than
`SIGQUIT` in this case, but it seems reasonable to respond to `SIGQUIT`
regardless.
Keeping the `mock` postgres cplane adaptor using "stock" tokio-postgres
allows us to remove a lot of dead weight from our actual postgres
connection logic.
## Problem
We saw a peculiar case where a pageserver apparently got a 0-tenant
response to `/re-attach` but we couldn't see the request landing on a
storage controller. It was hard to confirm retrospectively that the
pageserver was configured properly at the moment it sent the request.
## Summary of changes
- Log the URL to which we are sending the request
- Log the NodeId and metadata that we sent
## Problem
Sharded tenants should be run in a single AZ for best performance, so
that computes have AZ-local latency to all the shards.
Part of https://github.com/neondatabase/neon/issues/8264
## Summary of changes
- When we split a tenant, instead of updating each shard's preferred AZ
to wherever it is scheduled, propagate the preferred AZ from the parent.
- Drop the check in `test_shard_preferred_azs` that asserts shards end
up in their preferred AZ: this will not be true again until the
optimize_attachment logic is updated to make this so. The existing check
wasn't testing anything about scheduling, it was just asserting that we
set preferred AZ in a way that matches the way things happen to be
scheduled at time of split.
## Problem
In the batching PR
- https://github.com/neondatabase/neon/pull/9870
I stopped deducting the time-spent-in-throttle fro latency metrics,
i.e.,
- smgr latency metrics (`SmgrOpTimer`)
- basebackup latency (+scan latency, which I think is part of
basebackup).
The reason for stopping the deduction was that with the introduction of
batching, the trick with tracking time-spent-in-throttle inside
RequestContext and swap-replacing it from the `impl Drop for
SmgrOpTimer` no longer worked with >1 requests in a batch.
However, deducting time-spent-in-throttle is desirable because our
internal latency SLO definition does not account for throttling.
## Summary of changes
- Redefine throttling to be a page_service pagestream request throttle
instead of a throttle for repository `Key` reads through `Timeline::get`
/ `Timeline::get_vectored`.
- This means reads done by `basebackup` are no longer subject to any
throttle.
- The throttle applies after batching, before handling of the request.
- Drive-by fix: make throttle sensitive to cancellation.
- Rename metric label `kind` from `timeline_get` to `pagestream` to
reflect the new scope of throttling.
To avoid config format breakage, we leave the config field named
`timeline_get_throttle` and ignore the `task_kinds` field.
This will be cleaned up in a future PR.
## Trade-Offs
Ideally, we would apply the throttle before reading a request off the
connection, so that we queue the minimal amount of work inside the
process.
However, that's not possible because we need to do shard routing.
The redefinition of the throttle to limit pagestream request rate
instead of repository `Key` rate comes with several downsides:
- We're no longer able to use the throttle mechanism for other other
tasks, e.g. image layer creation.
However, in practice, we never used that capability anyways.
- We no longer throttle basebackup.
## Problem
`test_sharded_ingest` ingests a lot of data, which can cause shutdown to
be slow e.g. due to local "S3 uploads" or compactions. This can cause
test flakes during teardown.
Resolves#9740.
## Summary of changes
Perform an immediate shutdown of the cluster.
## Problem
We don't have good observability for memory usage. This would be useful
e.g. to debug OOM incidents or optimize performance or resource usage.
We would also like to use continuous profiling with e.g. [Grafana Cloud
Profiles](https://grafana.com/products/cloud/profiles-for-continuous-profiling/)
(see https://github.com/neondatabase/cloud/issues/14888).
This PR is intended as a proof of concept, to try it out in staging and
drive further discussions about profiling more broadly.
Touches https://github.com/neondatabase/neon/issues/9534.
Touches https://github.com/neondatabase/cloud/issues/14888.
Depends on #9779.
Depends on #9780.
## Summary of changes
Adds a HTTP route `/profile/heap` that takes a heap profile and returns
it. Query parameters:
* `format`: output format (`jemalloc` or `pprof`; default `pprof`).
Unlike CPU profiles (see #9764), heap profiles are not symbolized and
require the original binary to translate addresses to function names. To
make this work with Grafana, we'll probably have to symbolize the
process server-side -- this is left as future work, as is other output
formats like SVG.
Heap profiles don't work on macOS due to limitations in jemalloc.
## Problem
The extensions for Postgres v17 are ready but we do not test the
extensions shipped with v17
## Summary of changes
Build the test image based on Postgres v17. Run the tests for v17.
---------
Co-authored-by: Anastasia Lubennikova <anastasia@neon.tech>
This PR
- fixes smgr metrics https://github.com/neondatabase/neon/issues/9925
- adds an additional startup log line logging the current batching
config
- adds a histogram of batch sizes global and per-tenant
- adds a metric exposing the current batching config
The issue described #9925 is that before this PR, request latency was
only observed *after* batching.
This means that smgr latency metrics (most importantly getpage latency)
don't account for
- `wait_lsn` time
- time spent waiting for batch to fill up / the executor stage to pick
up the batch.
The fix is to use a per-request batching timer, like we did before the
initial batching PR.
We funnel those timers through the entire request lifecycle.
I noticed that even before the initial batching changes, we weren't
accounting for the time spent writing & flushing the response to the
wire.
This PR drive-by fixes that deficiency by dropping the timers at the
very end of processing the batch, i.e., after the `pgb.flush()` call.
I was **unable to maintain the behavior that we deduct
time-spent-in-throttle from various latency metrics.
The reason is that we're using a *single* counter in `RequestContext` to
track micros spent in throttle.
But there are *N* metrics timers in the batch, one per request.
As a consequence, the practice of consuming the counter in the drop
handler of each timer no longer works because all but the first timer
will encounter error `close() called on closed state`.
A failed attempt to maintain the current behavior can be found in
https://github.com/neondatabase/neon/pull/9951.
So, this PR remvoes the deduction behavior from all metrics.
I started a discussion on Slack about it the implications this has for
our internal SLO calculation:
https://neondb.slack.com/archives/C033RQ5SPDH/p1732910861704029
# Refs
- fixes https://github.com/neondatabase/neon/issues/9925
- sub-issue https://github.com/neondatabase/neon/issues/9377
- epic: https://github.com/neondatabase/neon/issues/9376
Before this PR, the storcon_cli didn't have a way to show the
tenant-wide information of the TenantDescribeResponse.
Sadly, the `Serialize` impl for the tenant config doesn't skip on
`None`, so, the output becomes a bit bloated.
Maybe we can use `skip_serializing_if(Option::is_none)` in the future.
=> https://github.com/neondatabase/neon/issues/9983
## Problem
I was touching `test_storage_controller_node_deletion` because for AZ
scheduling work I was adding a change to the storage controller (kick
secondaries during optimisation) that made a FIXME in this test defunct.
While looking at it I also realized that we can easily fix the way node
deletion currently doesn't use a proper ScheduleContext, using the
iterator type recently added for that purpose.
## Summary of changes
- A testing-only behavior in storage controller where if a secondary
location isn't yet ready during optimisation, it will be actively
polled.
- Remove workaround in `test_storage_controller_node_deletion` that
previously was needed because optimisation would get stuck on cold
secondaries.
- Update node deletion code to use a `TenantShardContextIterator` and
thereby a proper ScheduleContext
## Problem
After enabling LFC in tests and lowering `shared_buffers` we started
having more problems with `test_pg_regress`.
## Summary of changes
Set `shared_buffers` to 1MB to both exercise getPage requests/LFC, and
still have enough room for Postgres to operate. Everything smaller might
be not enough for Postgres under load, and can cause errors like 'no
unpinned buffers available'.
See Konstantin's comment [1] as well.
Fixes#9956
[1]:
https://github.com/neondatabase/neon/issues/9956#issuecomment-2511608097
On reconfigure, we no longer passed a port for the extension server
which caused us to not write out the neon.extension_server_port line.
Thus, Postgres thought we were setting the port to the default value of
0. PGC_POSTMASTER GUCs cannot be set at runtime, which causes the
following log messages:
> LOG: parameter "neon.extension_server_port" cannot be changed without
restarting the server
> LOG: configuration file
"/var/db/postgres/compute/pgdata/postgresql.conf" contains errors;
unaffected changes were applied
Fixes: https://github.com/neondatabase/neon/issues/9945
Signed-off-by: Tristan Partin <tristan@neon.tech>
The spec was written for the buggy protocol which we had before the one
more similar to Raft was implemented. Update the spec with what we
currently have.
ref https://github.com/neondatabase/neon/issues/8699
## Problem
The credentials providers tries to connect to AWS STS even when we use
plain Redis connections.
## Summary of changes
* Construct the CredentialsProvider only when needed ("irsa").
## Problem
`if: ${{ github.event.schedule }}` gets skipped if a previous step has
failed, but we want to run the step for both `success` and `failure`
## Summary of changes
- Add `!cancelled()` to notification step if-condition, to skip only
cancelled jobs
Fixes https://github.com/neondatabase/cloud/issues/20973.
This refactors `connect_raw` in order to return direct access to the
delayed notices.
I cannot find a way to test this with psycopg2 unfortunately, although
testing it with psql does return the expected results.
## Problem
We can't easily tell how far the state of shards is from their AZ
preferences. This can be a cause of performance issues, so it's
important for diagnosability that we can tell easily if there are
significant numbers of shards that aren't running in their preferred AZ.
Related: https://github.com/neondatabase/cloud/issues/15413
## Summary of changes
- In reconcile_all, count shards that are scheduled into the wrong AZ
(if they have a preference), and publish it as a prometheus gauge.
- Also calculate a statistic for how many shards wanted to reconcile but
couldn't.
This is clearly a lazy calculation: reconcile all only runs
periodically. But that's okay: shards in the wrong AZ is something that
only matters if it stays that way for some period of time.
Improves `wait_until` by:
* Use `timeout` instead of `iterations`. This allows changing the
timeout/interval parameters independently.
* Make `timeout` and `interval` optional (default 20s and 0.5s). Most
callers don't care.
* Only output status every 1s by default, and add optional
`status_interval` parameter.
* Remove `show_intermediate_error`, this was always emitted anyway.
Most callers have been updated to use the defaults, except where they
had good reason otherwise.
## Problem
We saw unexpected container terminations when running in k8s with with
small CPU resource requests.
The /status and /ready handlers called `maybe_forward`, which always
takes the lock on Service::inner.
If there is a lot of writer lock contention, and the container is
starved of CPU, this increases the likelihood that we will get killed by
the kubelet.
It isn't certain that this was a cause of issues, but it is a potential
source that we can eliminate.
## Summary of changes
- Revise logic to return immediately if the URL is in the non-forwarded
list, rather than calling maybe_forward
## Problem
See https://neondb.slack.com/archives/C04DGM6SMTM/p1732110190129479
We observe the following error in the logs
```
[XX000] ERROR: [NEON_SMGR] [shard 3] Incorrect prefetch read: status=1 response=0x7fafef335138 my=128 receive=128
```
most likely caused by changing `neon.readahead_buffer_size`
## Summary of changes
1. Copy shard state
2. Do not use prefetch_set_unused in readahead_buffer_resize
3. Change prefetch buffer overflow criteria
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
Current compute images for Postgres 14-16 don't build on Debian 12
because of issues with extensions.
This PR fixes that, but for the current setup, it is mostly a no-op
change.
## Summary of changes
- Use `/bin/bash -euo pipefail` as SHELL to fail earlier
- Fix `plv8` build: backport a trivial patch for v8
- Fix `postgis` build: depend `sfgal` version on Debian version instead
of Postgres version
Tested in: https://github.com/neondatabase/neon/pull/9849
#8564
## Problem
The main and backup consumption metric pushes are completely
independent,
resulting in different event time windows and different idempotency
keys.
## Summary of changes
* Merge the push tasks, but keep chunks the same size.
# Problem
The timeout-based batching adds latency to unbatchable workloads.
We can choose a short batching timeout (e.g. 10us) but that requires
high-resolution timers, which tokio doesn't have.
I thoroughly explored options to use OS timers (see
[this](https://github.com/neondatabase/neon/pull/9822) abandoned PR).
In short, it's not an attractive option because any timer implementation
adds non-trivial overheads.
# Solution
The insight is that, in the steady state of a batchable workload, the
time we spend in `get_vectored` will be hundreds of microseconds anyway.
If we prepare the next batch concurrently to `get_vectored`, we will
have a sizeable batch ready once `get_vectored` of the current batch is
done and do not need an explicit timeout.
This can be reasonably described as **pipelining of the protocol
handler**.
# Implementation
We model the sub-protocol handler for pagestream requests
(`handle_pagrequests`) as two futures that form a pipeline:
2. Batching: read requests from the connection and fill the current
batch
3. Execution: `take` the current batch, execute it using `get_vectored`,
and send the response.
The Reading and Batching stage are connected through a new type of
channel called `spsc_fold`.
See the long comment in the `handle_pagerequests_pipelined` for details.
# Changes
- Refactor `handle_pagerequests`
- separate functions for
- reading one protocol message; produces a `BatchedFeMessage` with just
one page request in it
- batching; tried to merge an incoming `BatchedFeMessage` into an
existing `BatchedFeMessage`; returns `None` on success and returns back
the incoming message in case merging isn't possible
- execution of a batched message
- unify the timeline handle acquisition & request span construction; it
now happen in the function that reads the protocol message
- Implement serial and pipelined model
- serial: what we had before any of the batching changes
- read one protocol message
- execute protocol messages
- pipelined: the design described above
- optionality for execution of the pipeline: either via concurrent
futures vs tokio tasks
- Pageserver config
- remove batching timeout field
- add ability to configure pipelining mode
- add ability to limit max batch size for pipelined configurations
(required for the rollout, cf
https://github.com/neondatabase/cloud/issues/20620 )
- ability to configure execution mode
- Tests
- remove `batch_timeout` parametrization
- rename `test_getpage_merge_smoke` to `test_throughput`
- add parametrization to test different max batch sizes and execution
moes
- rename `test_timer_precision` to `test_latency`
- rename the test case file to `test_page_service_batching.py`
- better descriptions of what the tests actually do
## On the holding The `TimelineHandle` in the pending batch
While batching, we hold the `TimelineHandle` in the pending batch.
Therefore, the timeline will not finish shutting down while we're
batching.
This is not a problem in practice because the concurrently ongoing
`get_vectored` call will fail quickly with an error indicating that the
timeline is shutting down.
This results in the Execution stage returning a `QueryError::Shutdown`,
which causes the pipeline / entire page service connection to shut down.
This drops all references to the
`Arc<Mutex<Option<Box<BatchedFeMessage>>>>` object, thereby dropping the
contained `TimelineHandle`s.
- => fixes https://github.com/neondatabase/neon/issues/9850
# Performance
Local run of the benchmarks, results in [this empty
commit](1cf5b1463f)
in the PR branch.
Key take-aways:
* `concurrent-futures` and `tasks` deliver identical `batching_factor`
* tail latency impact unknown, cf
https://github.com/neondatabase/neon/issues/9837
* `concurrent-futures` has higher throughput than `tasks` in all
workloads (=lower `time` metric)
* In unbatchable workloads, `concurrent-futures` has 5% higher
`CPU-per-throughput` than that of `tasks`, and 15% higher than that of
`serial`.
* In batchable-32 workload, `concurrent-futures` has 8% lower
`CPU-per-throughput` than that of `tasks` (comparison to tput of
`serial` is irrelevant)
* in unbatchable workloads, mean and tail latencies of
`concurrent-futures` is practically identical to `serial`, whereas
`tasks` adds 20-30us of overhead
Overall, `concurrent-futures` seems like a slightly more attractive
choice.
# Rollout
This change is disabled-by-default.
Rollout plan:
- https://github.com/neondatabase/cloud/issues/20620
# Refs
- epic: https://github.com/neondatabase/neon/issues/9376
- this sub-task: https://github.com/neondatabase/neon/issues/9377
- the abandoned attempt to improve batching timeout resolution:
https://github.com/neondatabase/neon/pull/9820
- closes https://github.com/neondatabase/neon/issues/9850
- fixes https://github.com/neondatabase/neon/issues/9835
## Problem
It appears that the Azure storage API tends to hang TCP connections more
than S3 does.
Currently we use a 2 minute timeout for all downloads. This is large
because sometimes the objects we download are large. However, waiting 2
minutes when doing something like downloading a manifest on tenant
attach is problematic, because when someone is doing a "create tenant,
create timeline" workflow, that 2 minutes is long enough for them
reasonably to give up creating that timeline.
Rather than propagate oversized timeouts further up the stack, we should
use a different timeout for objects that we expect to be small.
Closes: https://github.com/neondatabase/neon/issues/9836
## Summary of changes
- Add a `small_timeout` configuration attribute to remote storage,
defaulting to 30 seconds (still a very generous period to do something
like download an index)
- Add a DownloadKind parameter to DownloadOpts, so that callers can
indicate whether they expect the object to be small or large.
- In the azure client, use small timeout for HEAD requests, and for GET
requests if DownloadKind::Small is used.
- Use DownloadKind::Small for manifests, indices, and heatmap downloads.
This PR intentionally does not make the equivalent change to the S3
client, to reduce blast radius in case this has unexpected consequences
(we could accomplish the same thing by editing lots of configs, but just
skipping the code is simpler for right now)
## Problem
It was not always possible to judge what exactly some `cloud_admin`
connections were doing because we didn't consistently set
`application_name` everywhere.
## Summary of changes
Unify the way we connect to Postgres:
1. Switch to building configs everywhere
2. Always set `application_name` and make naming consistent
Follow-up for #9919
Part of neondatabase/cloud#20948
## Problem
To add Safekeeper heap profiling in #9778, we need to switch to an
allocator that supports it. Pageserver and proxy already use jemalloc.
Touches #9534.
## Summary of changes
Use jemalloc in Safekeeper.
## Problem
When picking locations for a shard, we should use a ScheduleContext that
includes all the other shards in the tenant, so that we apply proper
anti-affinity between shards. If we don't do this, then it can lead to
unstable scheduling, where we place a shard somewhere that the optimizer
will then immediately move it away from.
We didn't always do this, because it was a bit awkward to accumulate the
context for a tenant rather than just walking tenants.
This was a TODO in `handle_node_availability_transition`:
```
// TODO: populate a ScheduleContext including all shards in the same tenant_id (only matters
// for tenants without secondary locations: if they have a secondary location, then this
// schedule() call is just promoting an existing secondary)
```
This is a precursor to https://github.com/neondatabase/neon/issues/8264,
where the current imperfect scheduling during node evacuation hampers
testing.
## Summary of changes
- Add an iterator type that yields each shard along with a
schedulecontext that includes all the other shards from the same tenant
- Use the iterator to replace hand-crafted logic in optimize_all_plan
(functionally identical)
- Use the iterator in `handle_node_availability_transition` to apply
proper anti-affinity during node evacuation.
Our rust-postgres fork is getting messy. Mostly because proxy wants more
control over the raw protocol than tokio-postgres provides. As such,
it's diverging more and more. Storage and compute also make use of
rust-postgres, but in more normal usage, thus they don't need our crazy
changes.
Idea:
* proxy maintains their subset
* other teams use a minimal patch set against upstream rust-postgres
Reviewing this code will be difficult. To implement it, I
1. Copied tokio-postgres, postgres-protocol and postgres-types from
00940fcdb5
2. Updated their package names with the `2` suffix to make them compile
in the workspace.
3. Updated proxy to use those packages
4. Copied in the code from tokio-postgres-rustls 0.13 (with some patches
applied https://github.com/jbg/tokio-postgres-rustls/pull/32https://github.com/jbg/tokio-postgres-rustls/pull/33)
5. Removed as much dead code as I could find in the vendored libraries
6. Updated the tokio-postgres-rustls code to use our existing channel
binding implementation
Adds a benchmark for logical message WAL ingestion throughput
end-to-end. Logical messages are essentially noops, and thus ignored by
the Pageserver.
Example results from my MacBook, with fsync enabled:
```
postgres_ingest: 14.445 s
safekeeper_ingest: 29.948 s
pageserver_ingest: 30.013 s
pageserver_recover_ingest: 8.633 s
wal_written: 10,340 MB
message_count: 1310720 messages
postgres_throughput: 715 MB/s
safekeeper_throughput: 345 MB/s
pageserver_throughput: 344 MB/s
pageserver_recover_throughput: 1197 MB/s
```
See
https://github.com/neondatabase/neon/issues/9642#issuecomment-2475995205
for running analysis.
Touches #9642.
## Problem
We used `set_path()` to replace the database name in the connection
string. It automatically does url-safe encoding if the path is not
already encoded, but it does it as per the URL standard, which assumes
that tabs can be safely removed from the path without changing the
meaning of the URL. See, e.g.,
https://url.spec.whatwg.org/#concept-basic-url-parser. It also breaks
for DBs with properly %-encoded names, like with `%20`, as they are kept
intact, but actually should be escaped.
Yet, this is not true for Postgres, where it's completely valid to have
trailing tabs in the database name.
I think this is the PR that caused this regression
https://github.com/neondatabase/neon/pull/9717, as it switched from
`postgres::config::Config` back to `set_path()`.
This was fixed a while ago already [1], btw, I just haven't added a test
to catch this regression back then :(
## Summary of changes
This commit changes the code back to use
`postgres/tokio_postgres::Config` everywhere.
While on it, also do some changes around, as I had to touch this code:
1. Bump some logging from `debug` to `info` in the spec apply path. We
do not use `debug` in prod, and it was tricky to understand what was
going on with this bug in prod.
2. Refactor configuration concurrency calculation code so it was
reusable. Yet, still keep `1` in the case of reconfiguration. The
database can be actively used at this moment, so we cannot guarantee
that there will be enough spare connection slots, and the underlying
code won't handle connection errors properly.
3. Simplify the installed extensions code. It was spawning a blocking
task inside async function, which doesn't make much sense. Instead, just
have a main sync function and call it with `spawn_blocking` in the API
code -- the only place we need it to be async.
4. Add regression python test to cover this and related problems in the
future. Also, add more extensive testing of schema dump and DBs and
roles listing API.
[1]:
4d1e48f3b9
[2]:
https://www.postgresql.org/message-id/flat/20151023003445.931.91267%40wrigleys.postgresql.orgResolvesneondatabase/cloud#20869
## Problem
Currently, we rerun only known flaky tests. This approach was chosen to
reduce the number of tests that go unnoticed (by forcing people to take
a look at failed tests and rerun the job manually), but it has some
drawbacks:
- In PRs, people tend to push new changes without checking failed tests
(that's ok)
- In the main, tests are just restarted without checking
(understandable)
- Parametrised tests become flaky one by one, i.e. if `test[1]` is flaky
`, test[2]` is not marked as flaky automatically (which may or may not
be the case).
I suggest rerunning all failed tests to increase the stability of GitHub
jobs and using the Grafana Dashboard with flaky tests for deeper
analysis.
## Summary of changes
- Rerun all failed tests twice at max
## Problem
For the interpreted proto the pageserver is not returning the correct
LSN
in replies to keep alive requests. This is because the interpreted
protocol arm
was not updating `last_rec_lsn`.
## Summary of changes
* Return correct LSN in keep-alive responses
* Fix shard field in wal sender traces
We keep the practice of keeping the compiler up to date, pointing to the
latest release. This is done by many other projects in the Rust
ecosystem as well.
[Release notes](https://releases.rs/docs/1.83.0/).
Also update `cargo-hakari`, `cargo-deny`, `cargo-hack` and
`cargo-nextest` to their latest versions.
Prior update was in #9445.
## Problem
We currently see elevated levels of errors for GetBlob requests. This is
because 404 and 304 are counted as errors for metric reporting.
## Summary of Changes
Bring the implementation in line with the S3 client and treat 404 and
304 responses as ok for metric purposes.
Related: https://github.com/neondatabase/cloud/issues/20666
## Problem
For cancellation, a connection is open during all the cancel checks.
## Summary of changes
Spawn cancellation checks in the background, and close connection
immediately.
Use task_tracker for cancellation checks.
# Problem
VM (visibility map) pages are stored and managed as any regular relation
page, in the VM fork of the main relation. They are also sharded like
other pages. Regular WAL writes to the VM pages (typically performed by
vacuum) are routed to the correct shard as usual. However, VM pages are
also updated via `ClearVmBits` metadata records emitted when main
relation pages are updated. These metadata records were sent to all
shards, like other metadata records. This had the following effects:
* On shards responsible for VM pages, the `ClearVmBits` applies as
expected.
* On shard 0, which knows about the VM relation and its size but doesn't
necessarily have any VM pages, the `ClearVmBits` writes may have been
applied without also having applied the explicit WAL writes to VM pages.
* If VM pages are spread across multiple shards (unlikely with 256MB
stripe size), all shards may have applied `ClearVmBits` if the pages
fall within their local view of the relation size, even for pages they
do not own.
* On other shards, this caused a relation size cache miss and a DbDir
and RelDir lookup before dropping the `ClearVmBits`. With many
relations, this could cause significant CPU overhead.
This is not believed to be a correctness problem, but this will be
verified in #9914.
Resolves#9855.
# Changes
Route `ClearVmBits` metadata records only to the shards responsible for
the VM pages.
Verification of the current VM handling and cleanup of incomplete VM
pages on shard 0 (and potentially elsewhere) is left as follow-up work.
## Problem
close https://github.com/neondatabase/neon/issues/9859
## Summary of changes
Ensure that the deletion queue gets fully flushed (i.e., the deletion
lists get applied) during a graceful shutdown.
It is still possible that an incomplete shutdown would leave deletion
list behind and cause race upon the next startup, but we assume this
will unlikely happen, and even if it happened, the pageserver should
already be at a tainted state and the tenant should be moved to a new
tenant with a new generation number.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
* Promote two logs from mpsc send errors to error level. The channels
are unbounded and there shouldn't be errors.
* Fix one multiline log from anyhow::Error. Use Debug instead of
Display.
## Problem
When ingesting implicit `ClearVmBits` operations, we silently drop the
writes if the relation or page is unknown. There are implicit
assumptions around VM pages wrt. explicit/implicit updates, sharding,
and relation sizes, which can possibly drop writes incorrectly. Adding a
few metrics will allow us to investigate further and tighten up the
logic.
Touches #9855.
## Summary of changes
Add a `pageserver_wal_ingest_clear_vm_bits_unknown` metric to record
dropped `ClearVmBits` writes.
Also add comments clarifying the behavior of relation sizes on non-zero
shards.
Valid layer assumption is a necessary condition for a layer map to be
valid. It's a stronger check imposed by gc-compaction than the actual
valid layermap definition. Actually, the system can work as long as
there are no overlapping layer maps. Therefore, we degrade that into a
warning.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We don't have any observability for the relation size cache. We have
seen cache misses cause significant performance impact with high
relation counts.
Touches #9855.
## Summary of changes
Adds the following metrics:
* `pageserver_relsize_cache_entries`
* `pageserver_relsize_cache_hits`
* `pageserver_relsize_cache_misses`
* `pageserver_relsize_cache_misses_old`
## Problem
https://github.com/neondatabase/neon/pull/9746 lifted decoding and
interpretation of WAL to the safekeeper.
This reduced the ingested amount on the pageservers by around 10x for a
tenant with 8 shards, but doubled
the ingested amount for single sharded tenants.
Also, https://github.com/neondatabase/neon/pull/9746 uses bincode which
doesn't support schema evolution.
Technically the schema can be evolved, but it's very cumbersome.
## Summary of changes
This patch set addresses both problems by adding protobuf support for
the interpreted wal records and adding compression support. Compressed
protobuf reduced the ingested amount by 100x on the 32 shards
`test_sharded_ingest` case (compared to non-interpreted proto). For the
1 shard case the reduction is 5x.
Sister change to `rust-postgres` is
[here](https://github.com/neondatabase/rust-postgres/pull/33).
## Links
Related: https://github.com/neondatabase/neon/issues/9336
Epic: https://github.com/neondatabase/neon/issues/9329
## Problem
The `pre-merge-checks` workflow relies on the build-tools image.
If changes to the `build-tools` image have been merged into the main
branch since the last CI run for a PR (with other changes to the
`build-tools`), the image will be rebuilt during the merge queue run.
Otherwise, cached images are used.
Rebuilding the image adds approximately 10 minutes on x86-64 and 20
minutes on arm64 to the process.
## Summary of changes
- parametrise `build-build-tools-image` job with arch and Debian version
- Run `pre-merge-checks` only on Debian 12 x86-64 image
## Problem
ingest benchmark tests project migration to Neon involving steps
- COPY relation data
- create indexes
- create constraints
Previously we used only 4 copy jobs, 4 create index jobs and 7
maintenance workers. After increasing effective_io_concurrency on
compute we see that we can sustain more parallelism in the ingest bench
## Summary of changes
Increase copy jobs to 8, create index jobs to 8 and maintenance workers
to 16
## Problem
The RequestContext::span shouldn't live for the entire postgres
connection, only the handshake.
## Summary of changes
* Slight refactor to the RequestContext to discard the span upon
handshake completion.
* Make sure the temporary future for the handshake is dropped (not bound
to a variable)
* Runs our nightly fmt script
Before, we hardcoded the pg_version to 140000, while the code expected
version numbers like 14. Now we use an enum, and code from
`extension_server.rs` to auto-detect the correct version. The enum helps
when we add support for a version: enums ensure that compilation fails
if one forgets to put the version to one of the `match` locations.
cc https://github.com/neondatabase/neon/pull/9218
## Problem
Any errors from these async blocks are unconditionally logged at error
level
even though we already handle such errors based on context.
## Summary of changes
* Log raw errors from creating and executing cplane requests at debug
level.
* Inline macro calls to retain the correct callsite.
## Problem
The vast majority of the error/warn logs from cplane are about time or
data transfer quotas exceeded or endpoint-not-found errors and not
operational errors in proxy or cplane.
## Summary of changes
* Demote cplane error replies to info level.
* Raise other errors from warn back to error.
## Problem
For any given tenant shard, pageservers receive all of the tenant's WAL
from the safekeeper.
This soft-blocks us from using larger shard counts due to bandwidth
concerns and CPU overhead of filtering
out the records.
## Summary of changes
This PR lifts the decoding and interpretation of WAL from the pageserver
into the safekeeper.
A customised PG replication protocol is used where instead of sending
raw WAL, the safekeeper sends
filtered, interpreted records. The receiver drives the protocol
selection, so, on the pageserver side, usage
of the new protocol is gated by a new pageserver config:
`wal_receiver_protocol`.
More granularly the changes are:
1. Optionally inject the protocol and shard identity into the arguments
used for starting replication
2. On the safekeeper side, implement a new wal sending primitive which
decodes and interprets records
before sending them over
3. On the pageserver side, implement the ingestion of this new
replication message type. It's very similar
to what we already have for raw wal (minus decoding and interpreting).
## Notes
* This PR currently uses my [branch of
rust-postgres](https://github.com/neondatabase/rust-postgres/tree/vlad/interpreted-wal-record-replication-support)
which includes the deserialization logic for the new replication message
type. PR for that is open
[here](https://github.com/neondatabase/rust-postgres/pull/32).
* This PR contains changes for both pageservers and safekeepers. It's
safe to merge because the new protocol is disabled by default on the
pageserver side. We can gradually start enabling it in subsequent
releases.
* CI tests are running on https://github.com/neondatabase/neon/pull/9747
## Links
Related: https://github.com/neondatabase/neon/issues/9336
Epic: https://github.com/neondatabase/neon/issues/9329
## Problem
Prefetch is disabled at MacODS because `posix_fadvise` is not available.
But Neon prefetch is not using this function and for testing at MacOS is
it very convenient that prefetch is available.
## Summary of changes
Define `USE_PREFETCH` in Makefile.
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
We have a couple of CI workflows that still run on Debian Bullseye, and
the default Debian version in images is Bullseye as well (we explicitly
set building on Bookworm)
## Summary of changes
- Run `pgbench-pgvector` on Bookworm (fix a couple of packages)
- Run `trigger_bench_on_ec2_machine_in_eu_central_1` on Bookworm
- Change default `DEBIAN_VERSION` in Dockerfiles to Bookworm
- Make `pinned` docker tag an alias to `pinned-bookworm`
## Problem
close https://github.com/neondatabase/neon/issues/9761
The test assumed that no new L0 layers are flushed throughout the
process, which is not true.
## Summary of changes
Fix the test case `test_compaction_l0_memory` by flushing in-memory
layers before compaction.
Signed-off-by: Alex Chi Z <chi@neon.tech>
* The futures-util crate we use was yanked. Bump it and its siblings to
new patch release.
https://github.com/rust-lang/futures-rs/releases/tag/0.3.31
* cargo-deny: Drop an unused license.
* cargo-deny: Don't warn about duplicate crate. Duplicate crates are
unavoidable and the noise just hides real warnings.
## Problem
LFC is not enabled by default in tests, but it is enabled in production.
This increases the risk of errors in the production environment, which
were not found during the routine workflow.
However, enabling LFC for all the tests may overload the disk on our
servers and increase the number of failures.
So, we try enabling LFC in one case to evaluate the possible risk.
## Summary of changes
A new environment variable, USE_LFC is introduced. If it is set to true,
LFC is enabled by default in all the tests.
In our workflow, we enable LFC for PG17, release, x86-64, and disabled
for all other combinations.
---------
Co-authored-by: Alexey Masterov <alexeymasterov@neon.tech>
Co-authored-by: a-masterov <72613290+a-masterov@users.noreply.github.com>
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
Co-authored-by: Stas Kelvic <stas@neon.tech>
# Context
This PR contains PoC-level changes for a product feature that allows
onboarding large databases into Neon without going through the regular
data path.
# Changes
This internal RFC provides all the context
* https://github.com/neondatabase/cloud/pull/19799
In the language of the RFC, this PR covers
* the Importer code (`fast_import`)
* all the Pageserver changes (mgmt API changes, flow implementation,
etc)
* a basic test for the Pageserver changes
# Reviewing
As acknowledged in the RFC, the code added in this PR is not ready for
general availability.
Also, the **architecture is not to be discussed in this PR**, but in the
RFC and associated Slack channel instead.
Reviewers of this PR should take that into consideration.
The quality bar to apply during review depends on what area of the code
is being reviewed:
* Importer code (`fast_import`): practically anything goes
* Core flow (`flow.rs`):
* Malicious input data must be expected and the existing threat models
apply.
* The code must not be safe to execute on *dedicated* Pageserver
instances:
* This means in particular that tenants *on other* Pageserver instances
must not be affected negatively wrt data confidentiality, integrity or
availability.
* Other code: the usual quality bar
* Pay special attention to correct use of gate guards, timeline
cancellation in all places during shutdown & migration, etc.
* Consider the broader system impact; if you find potentially
problematic interactions with Storage features that were not covered in
the RFC, bring that up during the review.
I recommend submitting three separate reviews, for the three high-level
areas with different quality bars.
# References
(Internal-only)
* refs https://github.com/neondatabase/cloud/issues/17507
* refs https://github.com/neondatabase/company_projects/issues/293
* refs https://github.com/neondatabase/company_projects/issues/309
* refs https://github.com/neondatabase/cloud/issues/20646
---------
Co-authored-by: Stas Kelvich <stas.kelvich@gmail.com>
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
Co-authored-by: John Spray <john@neon.tech>
to keep it consistent with existing compute metrics.
flux-fleet change is not needed, because it doesn't have any filter by
metric name for compute metrics.
## Problem
Follow up of https://github.com/neondatabase/neon/pull/9682, that patch
didn't fully address the problem: what if shutdown fails due to whatever
reason and then we reattach the tenant? Then we will still remove the
future layer. The underlying problem is that the fix for #5878 gets
voided because of the generation optimizations.
Of course, we also need to ensure that delete happens after uploads, but
note that we only schedule deletes when there are no ongoing upload
tasks, so that's fine.
## Summary of changes
* Add a test case to reproduce the behavior (by changing the original
test case to attach the same generation).
* If layer upload happens after the deletion, drain the deletion queue
before uploading.
* If blocked_deletion is enabled, directly remove it from the
blocked_deletion queue.
* Local fs backend fix to avoid race between deletion and preload.
* test_emergency_mode does not need to wait for uploads (and it's
generally not possible to wait for uploads).
* ~~Optimize deletion executor to skip validation if there are no files
to delete.~~ this doesn't work
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Follow up to https://github.com/neondatabase/neon/pull/9682, hopefully
we can detect some issues or assure ourselves that this is ready for
production.
## Summary of changes
* Add a compaction-detach-ancestor smoke test.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
The HTTP router allowlists matched both on the path and the query
string. This meant that only `/profile/cpu` would be allowed without
auth, while `/profile/cpu?format=svg` would require auth.
Follows #9764.
## Summary of changes
* Match allowlists on URI path, rather than the entire URI.
* Fix the allowlist for Safekeeper to use `/profile/cpu` rather than the
old `/pprof/profile`.
* Just use a constant slice for the allowlist; it's only a handful of
items, and these handlers are not on hot paths.
## Problem
close https://github.com/neondatabase/neon/issues/9836
Looking at Azure SDK, the only related issue I can find is
https://github.com/azure/azure-sdk-for-rust/issues/1549. Azure uses
reqwest as the backend, so I assume there's some underlying magic
unknown to us that might have caused the stuck in #9836.
The observation is:
* We didn't get an explicit out of resource HTTP error from Azure.
* The connection simply gets stuck and times out.
* But when we retry after we reach the timeout, it succeeds.
This issue is hard to identify -- maybe something went wrong at the ABS
side, or something wrong with our side. But we know that a retry will
usually succeed if we give up the stuck connection.
Therefore, I propose the fix that we preempt stuck HTTP operation and
actively retry. This would mitigate the problem, while in the long run,
we need to keep an eye on ABS usage and see if we can fully resolve this
problem.
The reasoning of such timeout mechanism: we use a much smaller timeout
than before to preempt, while it is possible that a normal listing
operation would take a longer time than the initial timeout if it
contains a lot of keys. Therefore, after we terminate the connection, we
should double the timeout, so that such requests would eventually
succeed.
## Summary of changes
* Use exponential growth for ABS list timeout.
* Rather than using a fixed timeout, use a timeout that starts small and
grows
* Rather than exposing timeouts to the list_streaming caller as soon as
we see them, only do so after we have retried a few times
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Along with the migration to Python 3.11, I switched `C(str, Enum)` with
`C(StrEnum)`; one such example is the `PgVersion` enum.
It required more changes in `PgVersion` itself (before, it accepted both
`str` and `int`, and after it, it supports only `str`), which caused the
`test_bulk_insert` test to fail.
## Summary of changes
- `test_bulk_insert`: explicitly cast pg_version from `timeline_detail`
to str
I found the rightward drift of the `renew_jwks` function hard to review.
This PR splits out some major logic and uses early returns to make the
happy path more linear.
## Problem
We use a pretty old version of `mypy` 1.3 (released 1.5 years ago), it
produces false positives for `typing.Self`.
## Summary of changes
- Bump `mypy` from 1.3 to 1.13
- Fix new warnings and errors
- Use `typing.Self` whenever we `return self`
Without adding a newline, we can end up with a conf line that looks like
the following:
dynamic_shared_memory_type = mmap# Managed by compute_ctl: begin
This leads to Postgres logging:
LOG: configuration file
"/var/db/postgres/compute/pgdata/postgresql.conf" contains errors;
unaffected changes were applied
Signed-off-by: Tristan Partin <tristan@neon.tech>
The notifications need to be sent whenever the waiters heap changes, per
the comment in `update_status`. But if 'advance' is called when there
are no waiters, or the new LSN is lower than the waiters so that no one
needs to be woken up, there's no need to send notifications. This saves
some CPU cycles in the common case that there are no waiters.
## Problem
In https://github.com/neondatabase/neon/issues/9754 and the flakiness of
`test_readonly_node_gc`, we saw that although our logic for controlling
GC was sound, the validation of getpage requests was not, because it
could not consider LSN leases when requests arrived shortly after
restart.
Closes https://github.com/neondatabase/neon/issues/9754
## Summary of changes
This is the "Option 3" discussed verbally -- rather than holding back gc
cutoff, we waive the usual validation of request LSN if we are still
waiting for leases to be sent after startup
- When validating LSN in `wait_or_get_last_lsn`, skip the validation
relative to GC cutoff if the timeline is still in its LSN lease grace
period
- Re-enable test_readonly_node_gc
Calling unwrap on the encoder is a little overzealous. One of the errors
that can be returned by the encode function in particular is the
non-existence of metrics for a metric family, so we should prematurely
filter instances like that out. I believe that the cause of this panic
was caused by a race condition between the prometheus collector and the
compute collecting the installed extensions metric for the first time.
The HTTP server is spawned on a separate thread before we even start
bringing up Postgres.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
We don't have a convenient way to gather CPU profiles from a running
binary, e.g. during production incidents or end-to-end benchmarks, nor
during microbenchmarks (particularly on macOS).
We would also like to have continuous profiling in production, likely
using [Grafana Cloud
Profiles](https://grafana.com/products/cloud/profiles-for-continuous-profiling/).
We may choose to use either eBPF profiles or pprof profiles for this
(pending testing and discussion with SREs), but pprof profiles appear
useful regardless for the reasons listed above. See
https://github.com/neondatabase/cloud/issues/14888.
This PR is intended as a proof of concept, to try it out in staging and
drive further discussions about profiling more broadly.
Touches #9534.
Touches https://github.com/neondatabase/cloud/issues/14888.
## Summary of changes
Adds a HTTP route `/profile/cpu` that takes a CPU profile and returns
it. Defaults to a 5-second pprof Protobuf profile for use with e.g.
`pprof` or Grafana Alloy, but can also emit an SVG flamegraph. Query
parameters:
* `format`: output format (`pprof` or `svg`)
* `frequency`: sampling frequency in microseconds (default 100)
* `seconds`: number of seconds to profile (default 5)
Also integrates pprof profiles into Criterion benchmarks, such that
flamegraph reports can be taken with `cargo bench ... --profile-duration
<seconds>`. Output under `target/criterion/*/profile/flamegraph.svg`.
Example profiles:
* pprof profile (use [`pprof`](https://github.com/google/pprof)):
[profile.pb.gz](https://github.com/user-attachments/files/17756788/profile.pb.gz)
* Web interface: `pprof -http :6060 profile.pb.gz`
* Interactive flamegraph:
[profile.svg.gz](https://github.com/user-attachments/files/17756782/profile.svg.gz)
Fixes the masking for the CancelKeyData display format. Due to negative
i32 cast to u64, the top-bits all had `0xffffffff` prefix. On the
bitwise-or that followed, these took priority.
This PR also compresses 3 logs during sql-over-http into 1 log with
durations as label fields, as prior discussed.
## Problem
On Debian 12 (Bookworm), Python 3.11 is the latest available version.
## Summary of changes
- Update Python to 3.11 in build-tools
- Fix ruff check / format
- Fix mypy
- Use `StrEnum` instead of pair `str`, `Enum`
- Update docs
## Problem
I have made a mistake in merging Postgre PRs
## Summary of changes
Restore consistency of submodule referenced.
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
We saw a scale test failure when one shard went
secondary->attached->secondary in a short period of time -- the metrics
for the shard failed a validation assertion that is meant to ensure the
size metric matches the sum of layer sizes in the SecondaryDetail
struct.
This appears to be due to two SecondaryTenants being alive at the same
time -- the first one was shut down but still had its contributions to
the metrics.
Closes: https://github.com/neondatabase/neon/issues/9628
## Summary of changes
- Refactor code for validating metrics and call it in shutdown as well
as during downloads
- Move code for dropping per-tenant secondary metrics from drop() into
shutdown(), so that once shutdown() completes it is definitely safe to
instantiate another SecondaryTenant for the same tenant.
Before, `OpenTelemetry` errors were printed to stdout/stderr directly,
causing one of the few log lines without a timestamp, like:
```
OpenTelemetry trace error occurred. error sending request for url (http://localhost:4318/v1/traces)
```
Now, we print:
```
2024-11-21T02:24:20.511160Z INFO OpenTelemetry error: error sending request for url (http://localhost:4318/v1/traces)
```
I found this while investigating #9731.
## Problem
```
curl -H "Neon-Connection-String: postgresql://neondb_owner:PASSWORD@ep-autumn-rain-a58lubg0.us-east-2.aws.neon.tech/neondb?sslmode=require" https://ep-autumn-rain-a58lubg0.us-east-2.aws.neon.tech/sql -d '{"query":"SELECT 1","params":[]}'
```
For such a query, I also need to send `params`. Do I really need it?
## Summary of changes
I've marked `params` as optional
Adds support to the `find_garbage` command to restrict itself to a
partial tenant ID prefix, say `a`, and then it only traverses tenants
with IDs starting with `a`. One can now pass the `--tenant-id-prefix`
parameter.
That way, one can shard the `find_garbage` command and make it run in
parallel.
The PR also does a change of how `remote_storage` first removes trailing
`/`s, only to then add them in the listing function. It turns out that
this isn't neccessary and it prevents the prefix functionality from
working. S3 doesn't do this either.
## Problem
We were hitting this assertion in debug mode tests sometimes.
This case was being hit when the parent shard has no resident layers.
For instance, this is the case on split retry where the previous attempt
shut-down the parent and deleted local state for it. If the logical size
calculation does not download some layers before we get to the
hardlinking, then the assertion is hit.
## Summary of Changes
Remove the assertion. It's fine for the ancestor to not have any
resident layers at the time of the split.
Closes https://github.com/neondatabase/neon/issues/9412
Follow up to #9803
See https://github.com/neondatabase/cloud/issues/14378
In collaboration with @cloneable and @awarus, we sifted through logs and
simply demoted some logs to debug. This is not at all finished and there
are more logs to review, but we ran out of time in the session we
organised. In any slightly more nuanced cases, we didn't touch the log,
instead leaving a TODO comment.
I've also slightly refactored the sql-over-http body read/length reject
code. I can split that into a separate PR. It just felt natural after I
switched to `read_body_with_limit` as we discussed during the meet.
## Problem
Long ago, in #5299 the tenant states for migration are added, but
respected only in a coarse-grained way: when hinted not to do deletions,
tenants will just avoid doing all GC or compaction.
Skipping compaction is not necessary for AttachedMulti, as we will soon
become the primary attached location, and it is not a waste of resources
to proceed with compaction. Instead, per the RFC
https://github.com/neondatabase/neon/pull/5029/files), deletions should
be queued up in this state, and executed later when we switch to
AttachedSingle.
Avoiding compaction in AttachedMulti can have an operational impact if a
tenant is under significant write load, as a long-running migration can
result in a large accumulation of delta layers with commensurate impact
on read latency.
Closes: https://github.com/neondatabase/neon/issues/5396
## Summary of changes
- Add a 'config' part to RemoteTimelineClient so that it can be aware of
the mode of the tenant it belongs to, and wire this through for
construction + updates
- Add a special buffer for delayed deletions, and when in AttachedMulti
route deletions here instead of into the main remote client queue. This
is drained when transitioning to AttachedSingle. If the tenant is
detached or our process dies before then, then these objects are leaked.
- As a quality of life improvement, also use the remote timeline
client's knowledge of the tenant state to avoid submitting remote
consistent LSN updates for validation when in AttachedStale (as we know
these will fail)
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
## Problem
SLRU blocks, which can add up to several gigabytes, are currently
ingested by all shards, multiplying their capacity cost by the shard
count and slowing down ingest. We do this because all shards need the
SLRU pages to do timestamp->LSN lookup for GC.
Related: https://github.com/neondatabase/neon/issues/7512
## Summary of changes
- On non-zero shards, learn the GC offset from shard 0's index instead
of calculating it.
- Add a test `test_sharding_gc` that exercises this
- Do GC in test_pg_regress as a general smoke test that GC functions run
(e.g. this would fail if we were using SLRUs we didn't have)
In this PR we are still ingesting SLRUs everywhere, but not using them
any more. Part 2 PR (https://github.com/neondatabase/neon/pull/9786)
makes the change to not store them at all.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
## Problem
This test uses a gratuitous number of pageservers (16). This works fine
when there are plenty of system resources, but causes issues on test
runners that have limited resources and run many tests concurrently.
Related: https://github.com/neondatabase/neon/issues/9802
## Summary of changes
- Split from 2 shards to 4, instead of 4 to 8
- Don't give every shard a separate pageserver, let two locations share
each pageserver.
Net result is 4 pageservers instead of 16
## Problem
It is called context/ctx everywhere and the Monitoring suffix needlessly
confuses with proper monitoring code.
## Summary of changes
* Rename RequestMonitoring to RequestContext
* Rename RequestMonitoringInner to RequestContextInner
## Problem
I've noticed that we have 2 flaky tests which failed with error:
```
re.error: missing ), unterminated subpattern at position 21
```
- `test_timeline_archival_chaos` — has been already fixed
- `test_sharded_tad_interleaved_after_partial_success` — I didn't manage
to find the incorrect regex
[Internal link](https://neonprod.grafana.net/goto/yfmVHV7NR?orgId=1)
## Summary of changes
- Wrap `re.match` in `try..except` block and print incorrect regex
## Problem
We want to keep `#on-call-staging-stream` channel close to the prod one
and redirect notifications from failing benchmarks to another channel
for investigation.
## Summary of changes
- Send notifications regarding failures in `benchmarking` job to
`#on-call-staging-stream`
- Send notifications regarding failures in `periodic_pagebench` job to
`#on-call-staging-stream`
## Problem
Two recently observed log errors indicate safekeeper tasks for a
timeline running after that timeline's deletion has started.
- https://github.com/neondatabase/neon/issues/8972
- https://github.com/neondatabase/neon/issues/8974
These code paths do not have a mechanism that coordinates task shutdown
with the overall shutdown of the timeline.
## Summary of changes
- Add a `Gate` to `Timeline`
- Take the gate as part of resident timeline guard: any code that holds
a guard over a timeline staying resident should also hold a guard over
the timeline's total lifetime.
- Take the gate from the wal removal task
- Respect Timeline::cancel in WAL send/recv code, so that we do not
block shutdown indefinitely.
- Add a test that deletes timelines with open pageserver+compute
connections, to check these get torn down as expected.
There is some risk to introducing gates: if there is code holding a gate
which does not properly respect a cancellation token, it can cause
shutdown hangs. The risk of this for safekeepers is lower in practice
than it is for other services, because in a healthy timeline deletion,
the compute is shutdown first, then the timeline is deleted on the
pageserver, and finally it is deleted on the safekeepers -- that makes
it much less likely that some protocol handler will still be running.
Closes: #8972Closes: #8974
See https://github.com/neondatabase/cloud/issues/14378
In collaboration with @cloneable and @awarus, we sifted through logs and
simply demoted some logs to debug. This is not at all finished and there
are more logs to review, but we ran out of time in the session we
organised. In any slightly more nuanced cases, we didn't touch the log,
instead leaving a TODO comment.
In timeline preloading, we also do a preload for offloaded timelines.
This includes the download of `index-part.json`. Ultimately, such a
download is wasteful, therefore avoid it. Same goes for the remote
client, we just discard it immediately thereafter.
Part of #8088
---------
Co-authored-by: Christian Schwarz <christian@neon.tech>
Before, compute_ctl didn't have a good registry for what command would
run when, depending exclusively on sync code to apply changes. When
users have many databases/roles to manage, this step can take a
substantial amount of time, breaking assumptions about low (re)start
times in other systems.
This commit reduces the time compute_ctl takes to restart when changes
must be applied, by making all commands more or less blind writes, and
applying these commands in an asynchronous context, only waiting for
completion once we know the commands have all been sent.
Additionally, this reduces time spent by batching per-database
operations where previously we would create a new SQL connection for
every user-database operation we planned to execute.
## Problem
We have a bunch of duplicated code for automated releases. There will be
even more, once we have `release-compute` branch
(https://github.com/neondatabase/neon/pull/9637).
Another issue with the current `release` workflow is that it creates a
PR from the main as is. If we create 2 different releases from the
same commit, GitHub could mix up results from different PRs.
## Summary of changes
- Create a reusable workflow for releases
- Create an empty commit to differentiate releases
part of https://github.com/neondatabase/neon/issues/9114, we want to be
able to run partial gc-compaction in tests. In the future, we can also
expand this functionality to legacy compaction, so that we can trigger
compaction for a specific key range.
## Summary of changes
* Support passing compaction key range through pageserver routes.
* Refactor input parameters of compact related function to take the new
`CompactOptions`.
* Add tests for partial compaction. Note that the test may or may not
trigger compaction based on GC horizon. We need to improve the test case
to ensure things always get below the gc_horizon and the gc-compaction
can be triggered.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
close https://github.com/neondatabase/neon/issues/9730
The test case tests if anything goes wrong during pageserver restart +
*during timeline creation not complete*. Therefore, queue is stopped
error is normal in this case, except that it should be categorized as a
shutdown error instead of a real error.
## Summary of changes
* More comments for the test case.
* Queue stopped error will now be forwarded as
CreateTimelineError::ShuttingDown.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
The community decided to make a new off-schedule release due to ABI
breakage in last week's release. We're not affected by the ABI
breakage because we rebuild all extensions in our docker images, but
let's stay up-to-date. There were a few other fixes in the release
too.
Currently, local_proxy will write an error log if it doesn't find the
config file. This is expected for startup, so it's just noise. It is an
error if we do receive an explicit SIGHUP though.
I've also demoted the build info logs to be debug level. We don't need
them in the compute image since we have other ways to determine what
code is running.
Lastly, I've demoted SIGHUP signal handling from warn to info, since
it's not really a warning event.
See https://github.com/neondatabase/cloud/issues/10880 for more details
## Problem
The first version of the ingest benchmark had some parsing and reporting
logic in shell script inside GitHub workflow.
it is better to move that logic into a python testcase so that we can
also run it locally.
## Summary of changes
- Create new python testcase
- invoke pgcopydb inside python test case
- move the following logic into python testcase
- determine backpressure
- invoke pgcopydb and report its progress
- parse pgcopydb log and extract metrics
- insert metrics into perf test database
- add additional column to perf test database that can receive endpoint
ID used for pgcopydb run to have it available in grafana dashboard when
retrieving other metrics for an endpoint
## Example run
https://github.com/neondatabase/neon/actions/runs/11860622170/job/33056264386
Earlier work (#7547) has made the scrubber internally generic, but one
could only configure it to use S3 storage.
This is the final piece to make (most of, snapshotting still requires
S3) the scrubber be able to be configured via GenericRemoteStorage.
I.e. you can now set an env var like:
```
REMOTE_STORAGE_CONFIG='remote_storage = { bucket_name = "neon-dev-safekeeper-us-east-2d", bucket_region = "us-east-2" }
```
and the scrubber will read it instead.
There is a potential data corruption issue, not one I've encountered,
but it's still not hard to hit with some correct looking code given our
current architecture. It has to do with the timeline's memory object storage
via reference counted `Arc`s, and the removal of `retain_lsn` entries at
the drop of the last `Arc` reference.
The corruption steps are as follows:
1. timeline gets offloaded. timeline object A doesn't get dropped
though, because some long-running task accesses it
2. the same timeline gets unoffloaded again. timeline object B gets
created for it, timeline object A still referenced. both point to the
same timeline.
3. the task keeping the reference to timeline object A exits. destructor
for object A runs, removing `retain_lsn` in the timeline's parent.
4. the timeline's parent runs gc without the `retain_lsn` of the still
exant timleine's child, leading to data corruption.
In general we are susceptible each time when we recreate a `Timeline`
object in the same process, which happens both during a timeline
offload/unoffload cycle, as well as during an ancestor detach operation.
The solution this PR implements is to make the destructor for a timeline
as well as an offloaded timeline remove at most one `retain_lsn`.
PR #9760 has added a log line to print the refcounts at timeline
offload, but this only detects one of the places where we do such a
recycle operation. Plus it doesn't prevent the actual issue.
I doubt that this occurs in practice. It is more a defense in depth measure.
Usually I'd assume that the timeline gets dropped immediately in step 1,
as there is no background tasks referencing it after its shutdown.
But one never knows, and reducing the stakes of step 1 actually occurring
is a really good idea, from potential data corruption to waste of CPU time.
Part of #8088
## Problem
We don't take advantage of queue depth generated by the compute
on the pageserver. We can process getpage requests more efficiently
by batching them.
## Summary of changes
Batch up incoming getpage requests that arrive within a configurable
time window (`server_side_batch_timeout`).
Then process the entire batch via one `get_vectored` timeline operation.
By default, no merging takes place.
## Testing
* **Functional**: https://github.com/neondatabase/neon/pull/9792
* **Performance**: will be done in staging/pre-prod
# Refs
* https://github.com/neondatabase/neon/issues/9377
* https://github.com/neondatabase/neon/issues/9376
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
Tests that are marked with `run_only_on_default_postgres` do not run on
debug builds on CI because we run debug builds only for the latest
Postgres version (which is 17)
## Summary of changes
- Bump `PgVersion.DEFAULT` to `v17`
- Skip `test_timeline_archival_chaos` in debug builds
## Problem
We call `check-build-tools-image` twice for each workflow whenever we
use it, along with `build-build-tools-image`, once as a workflow itself,
and the second time from `build-build-tools-image`. This is not
necessary.
## Summary of changes
- Inline `check-build-tools-image` into `build-build-tools-image`
- Remove separate `check-build-tools-image` workflow
## Problem
Due to #9471 , the scale test occasionally gets 404s while trying to
modify the config of a timeline that belongs to a tenant being migrated.
We rarely see this narrow race in the field, but the test is quite good
at reproducing it.
## Summary of changes
- Ignore 404 errors in this test.
## Problem
See https://github.com/neondatabase/neon/issues/7750
test_wal_restore.sh is copying file to current working directory which
can cause interfere of test_wa_restore.py tests spawned of different
configurations.
## Summary of changes
Copy file to $DATA_DIR
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
The Postgres version was updated. The patch has to be updated
accordingly.
## Summary of changes
The patch of the regression test was updated.
## Problem
`no_sync` initially just skipped syncfs on startup (#9677). I'm also
interested in flaky tests that time out during pageserver shutdown while
flushing l0s, so to eliminate disk throughput as a source of issues
there,
## Summary of changes
- Drive-by change for test timeouts: add a couple more ::info logs
during pageserver startup so it's obvious which part got stuck.
- Add a SyncMode enum to configure VirtualFile and respect it in
sync_all and sync_data functions
- During pageserver startup, set SyncMode according to `no_sync`
## Problem
When processing pipelined `AppendRequest`s, we explicitly flush the WAL
every second and return an `AppendResponse`. However, the WAL is also
implicitly flushed on segment bounds, but this does not result in an
`AppendResponse`. Because of this, concurrent transactions may take up
to 1 second to commit and writes may take up to 1 second before sending
to the pageserver.
## Summary of changes
Advance `flush_lsn` when a WAL segment is closed and flushed, and emit
an `AppendResponse`. To accommodate this, track the `flush_lsn` in
addition to the `flush_record_lsn`.
## Problem
It turns out that `WalStreamDecoder::poll_decode` returns the start LSN
of the next record and not the end LSN of the current record. They are
not always equal. For example, they're not equal when the record in
question is an XLOG SWITCH record.
## Summary of changes
Rename things to reflect that.
In ea32f1d0a3, Matthias added a feature to
our extension to expose more granular wait events. However, due to the
typo, those wait events were never registered, so we used the more
generic wait events instead.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
We want to serialize interpreted records to send them over the wire from
safekeeper to pageserver.
## Summary of changes
Make `InterpretedWalRecord` ser/de. This is a temporary change to get
the bulk of the lift merged in
https://github.com/neondatabase/neon/pull/9746. For going to prod, we
don't want to use bincode since we can't evolve the schema.
Questions on serialization will be tackled separately.
PR #9308 has modified tenant activation code to take offloaded child
timelines into account for populating the list of `retain_lsn` values.
However, there is more places than just tenant activation where one
needs to update the `retain_lsn`s.
This PR fixes some bugs of the current code that could lead to
corruption in the worst case:
1. Deleting of an offloaded timeline would not get its `retain_lsn`
purged from its parent. With the patch we now do it, but as the parent
can be offloaded as well, the situatoin is a bit trickier than for
non-offloaded timelines which can just keep a pointer to their parent.
Here we can't keep a pointer because the parent might get offloaded,
then unoffloaded again, creating a dangling pointer situation. Keeping a
pointer to the *tenant* is not good either, because we might drop the
offloaded timeline in a context where a `offloaded_timelines` lock is
already held: so we don't want to acquire a lock in the drop code of
OffloadedTimeline.
2. Unoffloading a timeline would not get its `retain_lsn` values
populated, leading to it maybe garbage collecting values that its
children might need. We now call `initialize_gc_info` on the parent.
3. Offloading of a timeline would not get its `retain_lsn` values
registered as offloaded at the parent. So if we drop the `Timeline`
object, and its registration is removed, the parent would not have any
of the child's `retain_lsn`s around. Also, before, the `Timeline` object
would delete anything related to its timeline ID, now it only deletes
`retain_lsn`s that have `MaybeOffloaded::No` set.
Incorporates Chi's reproducer from #9753. cc
https://github.com/neondatabase/cloud/issues/20199
The `test_timeline_retain_lsn` test is extended:
1. it gains a new dimension, duplicating each mode, to either have the
"main" branch be the direct parent of the timeline we archive, or the
"test_archived_parent" branch intermediary, creating a three timeline
structure. This doesn't test anything fixed by this PR in particular,
just explores the vast space of possible configurations a little bit
more.
2. it gains two new modes, `offload-parent`, which tests the second
point, and `offload-no-restart` which tests the third point.
It's easy to verify the test actually is "sharp" by removing one of the
respective `self.initialize_gc_info()`, `gc_info.insert_child()` or
`ancestor_children.push()`.
Part of #8088
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Alex Chi Z <chi@neon.tech>
#9764, which adds profiling support to Safekeeper, pulls in the
dependency [`inferno`](https://crates.io/crates/inferno) via
[`pprof-rs`](https://crates.io/crates/pprof). This is licenced under the
[Common Development and Distribution License
1.0](https://spdx.org/licenses/CDDL-1.0.html), which is not allowed by
`cargo-deny`.
This patch allows the CDDL-1.0 license. It is a derivative of the
Mozilla Public License, which we already allow, but avoids some issues
around European copyright law that the MPL had. As such, I don't expect
this to be problematic.
## Problem
We didn't have a neat way to prevent auto-splitting of tenants. This
could be useful during incidents or for testing.
Closes https://github.com/neondatabase/neon/issues/9332
## Summary of changes
- Filter splitting candidates by scheduling policy
Analysis of the LR benchmarking tests indicates that in the duration of
test_subscriber_lag, a leftover 'slotter' replication slot can lead to
retained WAL growing on the publisher. This replication slot is not used
by any subscriber. The only purpose of the slot is to generate snapshot
files for the puspose of test_snap_files.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
After investigation, we think to make `test_readonly_node_gc` less
flaky, we need to make a proper fix (likely involving persisting part of
the lease state). See https://github.com/neondatabase/neon/issues/9754
for details.
## Summary of changes
- skip the test until proper fix.
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
## Problem
https://github.com/neondatabase/neon/issues/9240
## Summary of changes
Correctly truncate VM page instead just replacing it with zero page.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
## Problem
We are pining our fork of rust-postgres to a commit hash and that
prevents us from making
further changes to it. The latest commit in rust-postgres requires
https://github.com/neondatabase/neon/pull/8747,
but that seems to have gone stale. I reverted rust-postgres `neon`
branch to the pinned commit in
https://github.com/neondatabase/rust-postgres/pull/31.
## Summary of changes
Switch back to using the `neon` branch of the rust-postgres fork.
If WAL truncation fails in the middle it might leave some data on disk
above the write/flush LSN. In theory, concatenated with previous records
it might form bogus WAL (though very unlikely in practice because CRC
would protect from that). To protect from that, set
pending_wal_truncation flag: means before any WAL writes truncation must
be retried until it succeeds. We already did that in case of safekeeper
restart, now extend this mechanism for failures without restart. Also,
importantly, reset LSNs in the beginning of the operation, not in the
end, because once on disk deletion starts previous pointers are wrong.
All this most likely haven't created any problems in practice because
CRC protects from the consequences.
Tests for this are hard; simulation infrastructure might be useful here
in the future, but not yet.
## Problem
Historically, if a control component passed a pageserver "generation: 1"
this could be a quick way to corrupt a tenant by loading a historic
index.
Follows https://github.com/neondatabase/neon/pull/9383Closes#6951
## Summary of changes
- Introduce a Fatal variant to DownloadError, to enable index downloads
to signal when they have encountered a scary enough situation that we
shouldn't proceed to load the tenant.
- Handle this variant by putting the tenant into a broken state (no
matter which timeline within the tenant reported it)
- Add a test for this case
In the event that this behavior fires when we don't want it to, we have
ways to intervene:
- "Touch" an affected index to update its mtime (download+upload S3
object)
- If this behavior is triggered, it indicates we're attaching in some
old generation, so we should be able to fix that by manually bumping
generation numbers in the storage controller database (this should never
happen, but it's an option if it does)
and add /metrics endpoint to compute_ctl to expose such metrics
metric format example for extension pg_rag
with versions 1.2.3 and 1.4.2
installed in 3 and 1 databases respectively:
neon_extensions_installed{extension="pg_rag", version="1.2.3"} = 3
neon_extensions_installed{extension="pg_rag", version="1.4.2"} = 1
------
infra part: https://github.com/neondatabase/flux-fleet/pull/251
---------
Co-authored-by: Tristan Partin <tristan@neon.tech>
## Problem
Followup to https://github.com/neondatabase/neon/pull/9677 which enables
`no_sync` in tests. This can be merged once the next release has
happened.
## Summary of changes
- Always run pageserver with `no_sync = true` in tests.
psycopg2 has the following warning related to autocommit:
> By default, any query execution, including a simple SELECT will start
> a transaction: for long-running programs, if no further action is
> taken, the session will remain “idle in transaction”, an undesirable
> condition for several reasons (locks are held by the session, tables
> bloat…). For long lived scripts, either ensure to terminate a
> transaction as soon as possible or use an autocommit connection.
In the 2.9 release notes, psycopg2 also made the following change:
> `with connection` starts a transaction on autocommit transactions too
Some of these connections are indeed long-lived, so we were retaining
tons of WAL on the endpoints because we had a transaction pinned in the
past.
Link: https://www.psycopg.org/docs/news.html#what-s-new-in-psycopg-2-9
Link: https://github.com/psycopg/psycopg2/issues/941
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
`TimelinePersistentState::empty()`, used for tests and benchmarks, had a
hardcoded 16 MB WAL segment size. This caused confusion when attempting
to change the global segment size.
## Summary of changes
Inherit from `WAL_SEGMENT_SIZE` in `TimelinePersistentState::empty()`.
This GUC will drop replication slots if the size of the
pg_logical/snapshots directory (not including temp snapshot files)
becomes larger than the specified size. Keeping the size of this
directory smaller will help with basebackup size from the pageserver.
Part-of: https://github.com/neondatabase/neon/issues/8619
Signed-off-by: Tristan Partin <tristan@neon.tech>
The original value that we get is measured in microseconds. It comes
from a calculation using Postgres' GetCurrentTimestamp(), whihc is
implemented in terms of gettimeofday(2).
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
WAL segment fsyncs significantly affect WAL ingestion throughput.
`durable_rename()` is used when initializing every 16 MB segment, and
issues 3 fsyncs of which 1 was unnecessary.
## Summary of changes
Remove an fsync in `durable_rename` which is unnecessary with Linux and
ext4 (which we currently use). This improves WAL ingestion throughput by
up to 23% with large appends on my MacBook.
I had an impression that gc-compaction didn't test the case where the
first record of the key history is will_init because of there are some
code path that will panic in this case. Luckily it got fixed in
https://github.com/neondatabase/neon/pull/9026 so we can now implement
such tests.
Part of https://github.com/neondatabase/neon/issues/9114
## Summary of changes
* Randomly changed some images into will_init neon wal record
* Split `test_simple_bottom_most_compaction_deltas` into two test cases,
one of them has the bottom layer as delta layer with will_init flags,
while the other is the original one with image layers.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
The control file is flushed on the WAL ingest path when the commit LSN
advances by one segment, to bound the amount of recovery work in case of
a crash. This involves 3 additional fsyncs, which can have a significant
impact on WAL ingest throughput. This is to some extent mitigated by
`AppendResponse` not being emitted on segment bound flushes, since this
will prevent commit LSN advancement, which will be addressed separately.
## Summary of changes
Don't flush the control file on the WAL ingest path at all. Instead,
leave that responsibility to the timeline manager, but ask it to flush
eagerly if the control file lags the in-memory commit LSN by more than
one segment. This should not cause more than `REFRESH_INTERVAL` (300 ms)
additional latency before flushing the control file, which is
negligible.
Add a test that ensures the `retain_lsn` functionality works. Right now,
there is not a single test that is broken if offloaded or non-offloaded
timelines don't get registered at their parents, preventing gc from
discarding the ancestor_lsns of the children. This PR fills that gap.
The test has four modes:
* `offloaded`: offload the child timeline, run compaction on the parent
timeline, unarchive the child timeline, then try reading from it.
hopefully the data is still there.
* `offloaded-corrupted`: offload the child timeline, corrupts the
manifest in a way that the pageserver believes the timeline was
flattened. This is the closest we can get to pretend the `retain_lsn`
mechanism doesn't exist for offloaded timelines, so we can avoid adding
endpoints to the pageserver that do this manually for tests. The test
then checks that indeed data is corrupted and the endpoint can't be
started. That way we know that the test is actually working, and
actually tests the `retain_lsn` mechanism, instead of say the lsn lease
mechanism, or one of the many other mechanisms that impede gc.
* `archived`: the child timeline gets archived but doesn't get
offloaded. this currently matches the `None` case but we might have
refactors in the future that make archived timelines sufficiently
different from non-archived ones.
* `None`: the child timeline doesn't even get archived. this tests that
normal timelines participate in `retain_lsn`. I've made them locally not
participate in `retain_lsn` (via commenting out the respective
`ancestor_children.push` statement in tenant.rs) and ran the testsuite,
and not a single test failed. So this test is first of its kind.
Part of #8088.
This exporter logs an ERROR if a file called `postgres_exporter.yml` is
not located in its current working directory. We can silence it by
adding an empty config file and pointing the exporter at it.
Signed-off-by: Tristan Partin <tristan@neon.tech>
We found that exporting GH Workflow Runs in batch is more efficient due
to
- better utilisation of Github API
- and gh runners usage is rounded to minutes, so even when ad-hoc export
is done in 5-10 seconds, we billed for one minute usage
So now we introduce batch exporting, with version v0.2.x of github
workflow stats exporter.
How it's expected to work now:
- every 15 minutes we query for the workflow runs, created in last 2
hours
- to avoid missing workflows that ran for more than 2 hours, every night
(00:25) we will query workflows created in past 24 hours and export them
as well
- should we have query for even longer periods?
- lets see how it works with current schedule
- for longer periods like for days or weeks, it may require to adjust
logic and concurrency of querying data, so lets for now use simpler
version
The final patch for partial compaction, part of
https://github.com/neondatabase/neon/issues/9114, close
https://github.com/neondatabase/neon/issues/8921 (note that we didn't
implement parallel compaction or compaction scheduler for partial
compaction -- currently this needs to be scheduled by using a Python
script to split the keyspace, and in the future, automatically split
based on the key partitioning when the pageserver wants to trigger a
gc-compaction)
## Summary of changes
* Update the layer selection algorithm to use the same selection as full
compaction (everything intersect/below gc horizon)
* Update the layer selection algorithm to also generate a list of delta
layers that need to be rewritten
* Add the logic to rewrite delta layers and add them back to the layer
map
* Update test case to do partial compaction on deltas
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Removes some unnecessary initdb arguments, and fixes Neon for MacOS
since it doesn't seem to ship a C.UTF-8 locale.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
Running `pytest.skip(...)` in a test body instead of marking the test
with `@pytest.mark.skipif(...)` makes all fixtures to be initialised,
which is not necessary if the test is going to be skipped anyway.
Also, some tests are unnecessarily skipped (e.g. `test_layer_bloating`
on Postgres 17, or `test_idle_reconnections` at all) or run (e.g.
`test_parse_project_git_version_output_positive` more than on once
configuration) according to comments.
## Summary of changes
- Move `skip_on_postgres` / `xfail_on_postgres` /
`run_only_on_default_postgres` decorators to `fixture.utils`
- Add new `skip_in_debug_build` and `skip_on_ci` decorators
- Replace `pytest.skip(...)` calls with decorators where possible
## Problem
We have no specific benchmark testing project migration of postgresql
project with existing data into Neon.
Typical steps of such a project migration are
- schema creation in the neon project
- initial COPY of relations
- creation of indexes and constraints
- vacuum analyze
## Summary of changes
Add a periodic benchmark running 9 AM UTC every day.
In each run:
- copy a 200 GiB project that has realistic schema, data, tables,
indexes and constraints from another project into
- a new Neon project (7 CU fixed)
- an existing tenant, (but new branch and new database) that already has
4 TiB of data
- use pgcopydb tool to automate all steps and parallelize COPY and index
creation
- parse pgcopydb output and report performance metrics in Neon
performance test database
## Logs
This benchmark has been tested first manually and then as part of
benchmarking.yml workflow, example run see
https://github.com/neondatabase/neon/actions/runs/11757679870
## Problem
Once we enable the merge queue for the `main` branch, it won't be
possible to adjust the commit message right after pressing the "Squash
and merge" button and the PR title + description will be used as is.
To avoid extra noise in the commits in the `main` with the checklist
leftovers, I propose removing the checklist from the PR template and
keeping only the Problem / Summary of changes.
## Summary of changes
- Remove the checklist from the PR template
## Problem
We don't have a metric capturing the latency of segment initialization.
This can be significant due to fsyncs.
## Summary of changes
Add an `initialize_segment` variant of
`safekeeper_wal_storage_operation_seconds`.
Perf benchmarks produce a lot of layers.
## Summary of changes
Bumping the threshold and ignore the warning.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
GitHub API can return error 500, and it fails jobs that use
`actions/github-script` action.
## Summary of changes
- Add `retry: 500` to all `actions/github-script` usage
## Problem
We wish to stop using admin tokens in the infra repo, but step down
requests use the admin token.
## Summary of Changes
Introduce a new "ControllerPeer" scope and use it for step-down requests.
## Problem
The Merge queue doesn't work because it expects certain jobs, which we
don't have in the `pre-merge-checks` workflow.
But it turns out we can just create jobs/checks with the same names in
any workflow that we run.
## Summary of changes
- Add `conclusion` jobs
- Create `neon-cloud-e2e` status check
- Add a bunch of `if`s to handle cases with no relevant changes found
and prepare the workflow to run rust checks in the future
- List the workflow in `report-workflow-stats` to collect stats about it
It is possible at the point we shutdown the timeline, there are
still layer files we did not upload.
## Summary of changes
* If the queue is not empty, avoid offloading.
* Shutdown the timeline gracefully using the flush mode to
ensure all local files are uploaded before deleting the timeline
directory.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
We saw pageserver OOMs
https://github.com/neondatabase/cloud/issues/19715 for tenants doing
large writes. Add log lines around in-memory layers to hopefully collect
some info during my on-call shift next week.
## Summary of changes
* Estimate in-memory size of an in-mem layer.
* Print frozen layer number if there are too many layers accumulated in
memory.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Right now, our environments create databases with the C locale, which is
really unfortunate for users who have data stored in other languages
that they want to analyze. For instance, show_trgm on Hebrew text
currently doesn't work in staging or production.
I don't envision this being the final solution. I think this is just a
way to set a known value so the pageserver doesn't use its parent
environment. The final solution to me is exposing initdb parameters to
users in the console. Then they could use a different locale or encoding
if they so chose.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
To prevent breaking main after Python 3.11 PR get merged
we need to enable merge queue and run `check-codestyle-python`
job on it
## Summary of changes
- Move `check-codestyle-python` to a reusable workflow
- Run this workflow on `merge_group` event
In INC-317
https://neondb.slack.com/archives/C033RQ5SPDH/p1730815677932209, we saw
an interesting series of operations that would remove valid layer files
existing in the layer map.
* Timeline A starts compaction and generates an image layer Z but not
uploading it yet.
* Timeline B/C starts ancestor detaching (which should not affect
timeline A)
* The tenant gets restarted as part of the ancestor detaching process,
without increasing the generation number.
* Timeline A reloads, discovering the layer Z is a future layer, and
schedules a **deletion into the deletion queue**. This means that the
file will be deleted any time in the future.
* Timeline A starts compaction and generates layer Z again, adding it to
the layer map. Note that because we don't bump generation number during
ancestor detach, it has the same filename + generation number as the
original Z.
* Timeline A deletes layer Z from s3 + disk, and now we have a dangling
reference in the layer map, blocking all
compaction/logical_size_calculation process.
## Summary of changes
* We wait until all layers to be uploaded before shutting down the
tenants in `Flush` mode.
* Ancestor detach restarts now use this mode.
* Ancestor detach also waits for remote queue completion before starting
the detaching process.
* The patch ensures that we don't have any future image layer (or
something similar) after restart, but not fixing the underlying problem
around generation numbers.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Replication slots are now persisted using AUX files mechanism and
included in basebackup when replica is launched.
This slots are not somehow used at replica but hold WAL, which may cause
local disk space exhaustion.
## Summary of changes
Add `--replica` parameter to basebackup request and do not include
replication slot state files in basebackup for replica.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
In test environments, the `syncfs` that the pageserver does on startup
can take a long time, as other tests running concurrently might have
many gigabytes of dirty pages.
## Summary of changes
- Add a `no_sync` option to the pageserver's config.
- Skip syncfs on startup if this is set
- A subsequent PR (https://github.com/neondatabase/neon/pull/9678) will
enable this by default in tests. We need to wait until after the next
release to avoid breaking compat tests, which would fail if we set
no_sync & use an old pageserver binary.
Q: Why is this a different mechanism than safekeeper, which as a
--no-sync CLI?
A: Because the way we manage pageservers in neon_local depends on the
pageserver.toml containing the full configuration, whereas safekeepers
have a config file which is neon-local-specific and can drive a CLI
flag.
Q: Why is the option no_sync rather than sync?
A: For boolean configs with a dangerous value, it's preferable to make
"false" the safe option, so that any downstream future config tooling
that might have a "booleans are false by default" behavior (e.g. golang
structs) is safe by default.
Q: Why only skip the syncfs, and not all fsyncs?
A: Skipping all fsyncs would require more code changes, and the most
acute problem isn't fsyncs themselves (these just slow down a running
test), it's the syncfs (which makes a pageserver startup slow as a
result of _other_ tests)
set-docker-config-dir was replicated over multiple repositories.
The replica of this action was removed from this repository and it's
using the version from github.com/neondatabase/dev-actions instead
Compiling with nightly rust compiler, I'm getting a lot of errors like
this:
error: `if let` assigns a shorter lifetime since Edition 2024
--> proxy/src/auth/backend/jwt.rs:226:16
|
226 | if let Some(permit) = self.try_acquire_permit() {
| ^^^^^^^^^^^^^^^^^^^-------------------------
| |
| this value has a significant drop implementation which may observe a
major change in drop order and requires your discretion
|
= warning: this changes meaning in Rust 2024
= note: for more information, see issue #124085
<https://github.com/rust-lang/rust/issues/124085>
help: the value is now dropped here in Edition 2024
--> proxy/src/auth/backend/jwt.rs:241:13
|
241 | } else {
| ^
note: the lint level is defined here
--> proxy/src/lib.rs:8:5
|
8 | rust_2024_compatibility
| ^^^^^^^^^^^^^^^^^^^^^^^
= note: `#[deny(if_let_rescope)]` implied by
`#[deny(rust_2024_compatibility)]`
and this:
error: these values and local bindings have significant drop
implementation that will have a different drop order from that of
Edition 2021
--> proxy/src/auth/backend/jwt.rs:376:18
|
369 | let client = Client::builder()
| ------ these values have significant drop implementation and will
observe changes in drop order under Edition 2024
...
376 | map: DashMap::default(),
| ^^^^^^^^^^^^^^^^^^
|
= warning: this changes meaning in Rust 2024
= note: for more information, see issue #123739
<https://github.com/rust-lang/rust/issues/123739>
= note: `#[deny(tail_expr_drop_order)]` implied by
`#[deny(rust_2024_compatibility)]`
They are caused by the `rust_2024_compatibility` lint option.
When we actually switch to the 2024 edition, it makes sense to go
through all these and check that the drop order changes don't break
anything, but in the meanwhile, there's no easy way to avoid these
errors. Disable it, to allow compiling with nightly again.
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
build-tools image does not provide superuser, so additional packages can
not be installed during GitHub benchmarking workflows but need to be
added to the image
## Summary of changes
install pgcopydb version 0.17-1 or higher into build-tools bookworm
image
```bash
docker run -it neondatabase/build-tools:<tag>-bookworm-arm64 /bin/bash
...
nonroot@c23c6f4901ce:~$ LD_LIBRARY_PATH=/pgcopydb/lib /pgcopydb/bin/pgcopydb --version;
13:58:19.768 8 INFO Running pgcopydb version 0.17 from "/pgcopydb/bin/pgcopydb"
pgcopydb version 0.17
compiled with PostgreSQL 16.4 (Debian 16.4-1.pgdg120+2) on aarch64-unknown-linux-gnu, compiled by gcc (Debian 12.2.0-14) 12.2.0, 64-bit
compatible with Postgres 11, 12, 13, 14, 15, and 16
```
Example usage of that image in a workflow
https://github.com/neondatabase/neon/actions/runs/11725718371/job/32662681172#step:7:14
While setting up some tests, I noticed that we didn't support keycloak.
They make use of encryption JWKs as well as signature ones. Our current
jwks crate does not support parsing encryption keys which caused the
entire jwk set to fail to parse. Switching to lazy parsing fixes this.
Also while setting up tests, I couldn't use localhost jwks server as we
require HTTPS and we were using webpki so it was impossible to add a
custom CA. Enabling native roots addresses this possibility.
I saw some of our current e2e tests against our custom JWKS in s3 were
taking a while to fetch. I've added a timeout + retries to address this.
ref https://github.com/neondatabase/neon/issues/9441
The metrics from LR publisher testing project: ~300KB aux key deltas per
256MB files. Therefore, I think we can do compaction more aggressively
as these deltas are small and compaction can reduce layer download
latency. We also have a read path perf fix
https://github.com/neondatabase/neon/pull/9631 but I'd still combine the
read path fix with the reduce of the compaction threshold.
## Summary of changes
* reduce metadata compaction threshold
* use num of L1 delta layers as an indicator for metadata compaction
* dump more logs
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
When we create a new segment, we zero it out in order to avoid changing
the length and fsyncing metadata on every write. However, we zeroed it
out by writing 8 KB zero-pages, and Tokio file writes have non-trivial
overhead.
## Summary of changes
Zero out the segment using
[`File::set_len()`](https://docs.rs/tokio/latest/i686-unknown-linux-gnu/tokio/fs/struct.File.html#method.set_len)
instead. This will typically (depending on the filesystem) just write a
sparse file and omit the 16 MB of data entirely. This improves WAL
append throughput for large messages by over 400% with fsync disabled,
and 100% with fsync enabled.
## Problem
We don't have any benchmarks for Safekeeper WAL ingestion.
## Summary of changes
Add some basic benchmarks for WAL ingestion, specifically for
`SafeKeeper::process_msg()` (single append) and `WalAcceptor` (pipelined
batch ingestion). Also add some baseline file write benchmarks.
Fix direct reading from WAL buffers.
Pointer wasn't advanced which resulted in sending corrupted WAL if part
of read used WAL buffers and part read from the file. Also move it to
neon_walreader so that e.g. replication could also make use of it.
ref https://github.com/neondatabase/cloud/issues/19567
## Problem
Benchmarks need more control over the WAL generated by `WalGenerator`.
In particular, they need to vary the size of logical messages.
## Summary of changes
* Make `WalGenerator` generic over `RecordGenerator`, which constructs
WAL records.
* Add `LogicalMessageGenerator` which emits logical messages, with a
configurable payload.
* Minor tweaks and code reorganization.
There are no changes to the core logic or emitted WAL.
We need to use the shard associated with the layer file, not the shard
associated with our current tenant shard ID.
Due to shard splits, the shard IDs can refer to older files.
close https://github.com/neondatabase/neon/issues/9667
Fixes#9518.
## Problem
After removing the assertion `layers_removed == 0` in #9506, we could
miss breakage if we solely rely on the successful execution of the
`SELECT` query to check if lease is properly protecting layers. Details
listed in #9518.
Also, in integration tests, we sometimes run into the race condition
where getpage request comes before the lease get renewed (item 2 of
#8817), even if compute_ctl sends a lease renewal as soon as it sees a
`/configure` API calls that updates the `pageserver_connstr`. In this
case, we would observe a getpage request error stating that we `tried to
request a page version that was garbage collected` (as we seen in
[Allure
Report](https://neon-github-public-dev.s3.amazonaws.com/reports/pr-8613/11550393107/index.html#suites/3ccffb1d100105b98aed3dc19b717917/d1a1ba47bc180493)).
## Summary of changes
- Use layer map dump to verify if the lease protects what it claimed:
Record all historical layers that has `start_lsn <= lease_lsn` before
and after running timeline gc. This is the same check as
ad79f42460/pageserver/src/tenant/timeline.rs (L5025-L5027)
The set recorded after GC should contain every layer in the set recorded
before GC.
- Wait until log contains another successful lease request before
running the `SELECT` query after GC. We argued in #8817 that the bad
request can only exist within a short period after migration/restart,
and our test shows that as long as a lease renewal is done before the
first getpage request sent after reconfiguration, we will not have bad
request.
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
In https://github.com/neondatabase/neon/issues/9441, the tenant has a
lot of aux keys spread in multiple aux files. The perf tool shows that a
significant amount of time is spent on remove_overlapping_keys. For
sparse keyspaces, we don't need to report missing key errors anyways,
and it's very likely that we will need to read all layers intersecting
with the key range. Therefore, this patch adds a new fast path for
sparse keyspace reads that we do not track `unmapped_keyspace` in a
fine-grained way. We only modify it when we find an image layer.
In debug mode, it was ~5min to read the aux files for a dump of the
tenant, and now it's only 8s, that's a 60x speedup.
## Summary of changes
* Do not add sparse keys into `keys_done` so that remove_overlapping
does nothing.
* Allow `ValueReconstructSituation::Complete` to be updated again in
`ValuesReconstructState::update_key` for sparse keyspaces.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
I think I meant to make these changes over 6 months ago. alas, better
late than never.
1. should_reject doesn't eagerly intern the endpoint string
2. Rate limiter uses a std Mutex instead of a tokio Mutex.
3. Recently I introduced a `-local-proxy` endpoint suffix. I forgot to
add this to normalize.
4. Random but a small cleanup making the ControlPlaneEvent deser
directly to the interned strings.
## Problem
Pinning a tenant by setting Pause scheduling policy doesn't work because
drain/fill code moves the tenant around during deploys.
Closes: https://github.com/neondatabase/neon/issues/9612
## Summary of changes
- In drain, only move a tenant if it is in Active or Essential mode
- In fill, only move a tenant if it is in Active mode.
The asymmetry is a bit annoying, but it faithfully respects the purposes
of the modes: Essential is meant to endeavor to keep the tenant
available, which means it needs to be drained but doesn't need to be
migrated during fills.
## Problem
https://github.com/neondatabase/neon/pull/9524 split the decoding and
interpretation step from ingestion.
The output of the first phase is a `wal_decoder::models::InterpretedWalRecord`.
Before this patch set that struct contained a list of `Value` instances.
We wish to lift the decoding and interpretation step to the safekeeper,
but it would be nice if the safekeeper gave us a batch containing the raw data instead of actual values.
## Summary of changes
Main goal here is to make `InterpretedWalRecord` hold a raw buffer which
contains pre-serialized Values.
For this we do:
1. Add a `SerializedValueBatch` type. This is `inmemory_layer::SerializedBatch` with some
extra functionality for extension, observing values for shard 0 and tests.
2. Replace `inmemory_layer::SerializedBatch` with `SerializedValueBatch`
3. Make `DatadirModification` maintain a `SerializedValueBatch`.
### `DatadirModification` changes
`DatadirModification` now maintains a `SerializedValueBatch` and extends
it as new WAL records come in (to avoid flushing to disk on every
record).
In turn, this cascaded into a number of modifications to
`DatadirModification`:
1. Replace `pending_data_pages` and `pending_zero_data_pages` with `pending_data_batch`.
2. Removal of `pending_zero_data_pages` and its cousin `on_wal_record_end`
3. Rename `pending_bytes` to `pending_metadata_bytes` since this is what it tracks now.
4. Adapting of various utility methods like `len`, `approx_pending_bytes` and `has_dirty_data_pages`.
Removal of `pending_zero_data_pages` and the optimisation associated
with it ((1) and (2)) deserves more detail.
Previously all zero data pages went through `pending_zero_data_pages`.
We wrote zero data pages when filling gaps caused by relation extension
(case A) and when handling special wal records (case B). If it happened
that the same WAL record contained a non zero write for an entry in
`pending_zero_data_pages` we skipped the zero write.
Case A: We handle this differently now. When ingesting the
`SerialiezdValueBatch` associated with one PG WAL record, we identify the gaps and fill the
them in one go. Essentially, we move from a per key process (gaps were filled after each
new key), and replace it with a per record process. Hence, the optimisation is not
required anymore.
Case B: When the handling of a special record needs to zero out a key,
it just adds that to the current batch. I inspected the code, and I
don't think the optimisation kicked in here.
The overall idea of the PR is to rename a few types to make their
purpose more clear, reduce abstraction where not needed, and move types
to to more better suited modules.
The PROXY Protocol V2 offers a "command" concept. It can be of two
different values. "Local" and "Proxy". The spec suggests that "Local" be
used for health-checks. We can thus use this to silence logging for such
health checks such as those from NLB.
This additionally refactors the flow to be a bit more type-safe, self
documenting and using zerocopy deser.
## Problem
While experimenting with `MAX_SEND_SIZE` for benchmarking, I saw stack
overflows when increasing it to 1 MB. Turns out a few buffers of this
size are stack-allocated rather than heap-allocated. Even at the default
128 KB size, that's a bit large to allocate on the stack.
## Summary of changes
Heap-allocate buffers of size `MAX_SEND_SIZE`.
Unify client, EndpointConnPool and DbUserConnPool for remote and local
conn.
- Use new ClientDataEnum for additional client data.
- Add ClientInnerCommon client structure.
- Remove Client and EndpointConnPool code from local_conn_pool.rs
## Problem
First issues noticed when trying to run scrubber find-garbage on Azure:
- Azure staging contains projects with -1 set for max_project_size:
apparently the control plane treats this as a signed field.
- Scrubber code assumed that listing projects should filter to
aws-$REGION. This is no longer needed (per comment in the code) because
we know hit region-local APIs.
This PR doesn't make it work all the way (`init_remote` still assumes
S3), but these are necessary precursors.
## Summary of changes
- Change max-project_size from unsigned to signed
- Remove region filtering in favor of simply using the right region's
API (which we already do)
Since 5f83c9290b482dc90006c400dfc68e85a17af785/#1504 we've had
duplication in construction of models::TenantConfig, where both
constructs contained the same code. This PR removes one of the two
locations to avoid the duplication.
## Problem
Tenant operations may return `409 Conflict` if the tenant is shutting
down. This status code is not retried by the control plane, causing
user-facing errors during pageserver restarts. Operations should instead
return `503 Service Unavailable`, which may be retried for idempotent
operations.
## Summary of changes
Convert
`GetActiveTenantError::WillNotBecomeActive(TenantState::Stopping)` to
`ApiError::ShuttingDown` rather than `ApiError::Conflict`. This error is
returned by `Tenant::wait_to_become_active` in most (all?)
tenant/timeline-related HTTP routes.
* Also rename `AuthFailed` variant to `PasswordFailed`.
* Before this all JWT errors end up in `AuthError::AuthFailed()`,
expects a username and also causes cache invalidation.
Problem
-------
Tests that directly call the Pageserver Management API to set tenant
config are flaky if the Pageserver is managed by Storcon because Storcon
is the source of truth and may (theoretically) reconcile a tenant at any
time.
Solution
--------
Switch all users of
`set_tenant_config`/`patch_tenant_config_client_side`
to use the `env.storage_controller.pageserver_api()`
Future Work
-----------
Prevent regressions from creeping in.
And generally clean up up tenant configuration.
Maybe we can avoid the Pageserver having a default tenant config at all
and put the default into Storcon instead?
* => https://github.com/neondatabase/neon/issues/9621
Refs
----
fixes https://github.com/neondatabase/neon/issues/9522
## Problem
We don't have any observability for Safekeeper WAL receiver queues.
## Summary of changes
Adds a few WAL receiver metrics:
* `safekeeper_wal_receivers`: gauge of currently connected WAL
receivers.
* `safekeeper_wal_receiver_queue_depth`: histogram of queue depths per
receiver, sampled every 5 seconds.
* `safekeeper_wal_receiver_queue_depth_total`: gauge of total queued
messages across all receivers.
* `safekeeper_wal_receiver_queue_size_total`: gauge of total queued
message sizes across all receivers.
There are already metrics for ingested WAL volume: `written_wal_bytes`
counter per timeline, and `safekeeper_write_wal_bytes` per-request
histogram.
AWS/azure private link shares extra information in the "TLV" values of
the proxy protocol v2 header. This code doesn't action on it, but it
parses it as appropriate.
It seems the ecosystem is not so keen on moving to aws-lc-rs as it's
build setup is more complicated than ring (requiring cmake).
Eventually I expect the ecosystem should pivot to
https://github.com/ctz/graviola/tree/main/rustls-graviola as it
stabilises (it has a very simply build step and license), but for now
let's try not have a headache of juggling two crypto libs.
I also noticed that tonic will just fail with tls without a default
provider, so I added some defensive code for that.
## Problem
In https://github.com/neondatabase/neon/pull/9589, timeline offload code
is modified to return an explicit error type rather than propagating
anyhow::Error. One of the 'Other' cases there is I/O errors from local
timeline deletion, which shouldn't need to exist, because our policy is
not to try and continue running if the local disk gives us errors.
## Summary of changes
- Make `delete_local_timeline_directory` and use `.fatal_err(` on I/O
errors
---------
Co-authored-by: Erik Grinaker <erik@neon.tech>
Adds a Python benchmark for sharded ingestion. This ingests 7 GB of WAL
(100M rows) into a Safekeeper and fans out to 10 shards running on 10
different pageservers. The ingest volume and duration is recorded.
## Problem
https://neondb.slack.com/archives/C04DGM6SMTM/p1727872045252899
See https://github.com/neondatabase/neon/issues/9240
## Summary of changes
Add `!page_is_new` check before assigning page lsn.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
We may sometimes use scheduling modes like `Pause` to pin a tenant in
its current location for operational reasons. It is undesirable for the
chaos task to make any changes to such projects.
## Summary of changes
- Add a check for scheduling mode
- Add a log line when we do choose to do a chaos action for a tenant:
this will help us understand which operations originate from the chaos
task.
## Problem
We don't have any observability into full WalAcceptor queues per
timeline.
## Summary of changes
Logs a message when a WalAcceptor send has blocked for 5 seconds, and
another message when the send completes. This implies that the log
frequency is at most once every 5 seconds per timeline, so we don't need
further throttling.
## Problem
The IAM role associated with our github action runner supports a max
token expiration which is lower than the value we tried.
## Summary of changes
Since we believe to have understood the performance regression we (by
ensuring availability zone affinity of compute and pageserver) the job
should again run in lower than 5 hours and we revert this change instead
of increasing the max session token expiration in the IAM role which
would reduce our security.
## Problem
`tenant_get_shards()` does not work with a sharded tenant on 1
pageserver, as it assumes an unsharded tenant in this case. This special
case appears to have been added to handle e.g. `test_emergency_mode`,
where the storage controller is stopped. This breaks e.g. the sharded
ingest benchmark in #9591 when run with a single shard.
## Summary of changes
Correctly look up shards even with a single pageserver, but add a
special case that assumes an unsharded tenant if the storage controller
is stopped and the caller provides an explicit pageserver, in order to
accomodate `test_emergency_mode`.
## Problem
We wish for the deployment orchestrator to use infra scoped tokens,
but storcon endpoints it's using require admin scoped tokens.
## Summary of Changes
Switch over all endpoints that are used by the deployment orchestrator
to use an infra scoped token. This causes no breakage during mixed
version scenarios because admin scoped tokens allow access to all
endpoints. The deployment orchestrator can cut over to the infra token
after this commit touches down in prod.
Once this commit is released we should also update the tests code to use
infra scoped tokens where appropriate. Currently it would fail on the
[compat tests](9761b6a64e/test_runner/regress/test_storage_controller.py (L69-L71)).
## Problem
The final part of https://github.com/neondatabase/neon/issues/9543 will
be a chaos test that creates/deletes/archives/offloads timelines while
restarting pageservers and migrating tenants. Developing that test
showed up a few places where we log errors during normal shutdown.
## Summary of changes
- UninitializedTimeline's drop should log at info severity: this is a
normal code path when some part of timeline creation encounters a
cancellation `?` path.
- When offloading and finding a `RemoteTimelineClient` in a
non-initialized state, this is not an error and should not be logged as
such.
- The `offload_timeline` function returned an anyhow error, so callers
couldn't gracefully pick out cancellation errors from real errors:
update this to have a structured error type and use it throughout.
## Problem
clickbench regression causes clickbench to run >9 hours and the AWS
session token is expired before the run completes
## Summary of changes
extend lifetime of session token for this job to 12 hours
## Problem
Decoding and ingestion are still coupled in `pageserver::WalIngest`.
## Summary of changes
A new type is added to `wal_decoder::models`, InterpretedWalRecord. This
type contains everything that the pageserver requires in order to ingest
a WAL record. The highlights are the `metadata_record` which is an
optional special record type to be handled and `blocks` which stores
key, value pairs to be persisted to storage.
This type is produced by
`wal_decoder::models::InterpretedWalRecord::from_bytes` from a raw PG
wal record.
The rest of this commit separates decoding and interpretation of the PG
WAL record from its application in `WalIngest::ingest_record`.
Related: https://github.com/neondatabase/neon/issues/9335
Epic: https://github.com/neondatabase/neon/issues/9329
If we delete a timeline that has childen, those children will have their
data corrupted. Therefore, extend the already existing safety check to
offloaded timelines as well.
Part of #8088
In https://github.com/neondatabase/neon/issues/9032, I would like to
eventually add a `generation` field to the consumption metrics cache.
The current encoding is not backward compatible and it is hard to add
another field into the cache. Therefore, this patch refactors the format
to store "field -> value", and it's easier to maintain backward/forward
compatibility with this new format.
## Summary of changes
* Add `NewRawMetric` as the new format.
* Add upgrade path. When opening the disk cache, the codepath first
inspects the `version` field, and decide how to decode.
* Refactor metrics generation code and tests.
* Add tests on upgrade / compatibility with the old format.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Constructing a remote client is no big deal. Yes, it means an extra
download from S3 but it's not that expensive. This simplifies code paths
and scenarios to test. This unifies timelines that have been recently
offloaded with timelines that have been offloaded in an earlier
invocation of the process.
Part of #8088
Disallow a request for timeline ancestor detach if either the to be
detached timeline, or any of the to be reparented timelines are
offloaded or archived.
In theory we could support timelines that are archived but not
offloaded, but archived timelines are at the risk of being offloaded, so
we treat them like offloaded timelines. As for offloaded timelines, any
code to "support" them would amount to unoffloading them, at which point
we can just demand to have the timelines be unarchived.
Part of #8088
This will tell us how much time the compute has spent throttled if
pageserver/safekeeper cannot keep up with WAL generation.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
We don't have a convenient way to generate WAL records for benchmarks
and tests.
## Summary of changes
Adds a WAL generator, exposed as an iterator. It currently only
generates logical messages (noops), but will be extended to write actual
table rows later.
Some existing code for WAL generation has been replaced with this
generator, to reduce duplication.
In July of 2023, Bojan and Chi authored
92aee7e07f. Our in production pageservers
are most definitely at a version where they all support gzipped
basebackups.
## Problem
When tenant manifest objects are written without a generation suffix,
concurrently attached pageservers may stamp on each others writes of the
manifest and cause undefined behavior.
Closes: #9543
## Summary of changes
- Use download_generation_object helper when reading manifests, to
search for the most recent generation
- Use Tenant::generation as the generation suffix when writing
manifests.
This patch contains various improvements for the pagectl tool.
## Summary of changes
* Rewrite layer name parsing: LayerName now supports all variants we use
now.
* Drop pagectl's own layer parsing function, use LayerName in the
pageserver crate.
* Support image layer dumping in the layer dump command using
ImageLayer::dump, drop the original implementation.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
See https://github.com/neondatabase/neon/pull/9458
This PR separates PS related changes in #9458 from compute_ctl changes
to enforce that PS is deployed before compute.
## Summary of changes
This PR adds handlings of `--replica` parameters of backebackup to page
server.
## Checklist before requesting a review
- [ ] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
---------
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
## Problem
Uploads of the tenant manifest could race between different tasks,
resulting in unexpected results in remote storage.
Closes: https://github.com/neondatabase/neon/issues/9556
## Summary of changes
- Create a central function for uploads that takes a tokio::sync::Mutex
- Store the latest upload in that Mutex, so that when there is lots of
concurrency (e.g. archive 20 timelines at once) we can coalesce their
manifest writes somewhat.
## Problem
Indices used to be the only kind of object where we had to search across
generations to find the most recent one. As of
https://github.com/neondatabase/neon/issues/9543, manifests will need
the same treatment.
## Summary of changes
- Refactor download_index_part to a generic download_generation_object
function, which will be usable for downloading manifest objects as well.
python based regression test setup for auth_broker. This uses a http
mock for cplane as well as the JWKs url.
complications:
1. We cannot just use local_proxy binary, as that requires the
pg_session_jwt extension which we don't have available in the current
test suite
2. We cannot use just any old http mock for local_proxy, as auth_broker
requires http2 to local_proxy
as such, I used the h2 library to implement an echo server - copied from
the examples in the h2 docs.
In the base64 payload of an aws cognito jwt, I saw the following:
```
"iss":"https:\/\/cognito-idp.us-west-2.amazonaws.com\/us-west-2_redacted"
```
issuers are supposed to be URLs, and URLs are always valid un-escaped
JSON. However, `\/` is a valid escape character so what AWS is doing is
technically correct... sigh...
This PR refactors the test suite and adds a new regression test for
cognito.
## Problem
click bench job in benchmarking workflow has a performance regression
causing it to run in timeout of max job run.
Suspected root cause:
Project has been migrated from single pageserver to storage controller
managed project on Oct 14th.
Since then the regression shows.
## Summary of changes
Increase timeout of pytest to 12 hours.
Increase job timeout to 12 hours
As pointed out in
https://github.com/neondatabase/neon/pull/9489#discussion_r1814699683 ,
we currently didn't support deletion for offloaded timelines after the
timeline has been loaded from the manifest instead of having been
offloaded.
This was because the upload queue hasn't been initialized yet. This PR
thus initializes the timeline and shuts it down immediately.
Part of #8088
## Problem
We wish to have high level WAL decoding logic in `wal_decoder::decoder`
module.
## Summary of Changes
For this we need the `Value` and `NeonWalRecord` types accessible there, so:
1. Move `Value` and `NeonWalRecord` to `pageserver::value` and
`pageserver::record` respectively.
2. Get rid of `pageserver::repository` (follow up from (1))
3. Move PG specific WAL record types to `postgres_ffi::walrecord`. In
theory they could live in `wal_decoder`, but it would create a circular
dependency between `wal_decoder` and `postgres_ffi`. Long term it makes
sense for those types to be PG version specific, so that will work out nicely.
4. Move higher level WAL record types (to be ingested by pageserver)
into `wal_decoder::models`
Related: https://github.com/neondatabase/neon/issues/9335
Epic: https://github.com/neondatabase/neon/issues/9329
Currently, all callers of `unoffload_timeline` ensure that the tenant
the unoffload operation is called on is active. We rely on it being
active as we activate the timeline below and don't want to race with the
activation code of the tenant (in the worst case, activating a timeline
twice).
Therefore, add this assertion.
Part of #8088
We will only have a C string if the specified role is a string.
Otherwise, we need to resolve references to public, current_role,
current_user, and session_user.
Fixes: https://github.com/neondatabase/cloud/issues/19323
Signed-off-by: Tristan Partin <tristan@neon.tech>
As a DBaaS provider, Neon needs to provide a stable platform for
customers to build applications upon. At the same time however, we also
need to enable customers to use the latest and greatest technology, so
they can prototype their work, and we can solicit feedback. If all
extensions are treated the same in terms of stability, it is hard to
meet that goal.
There are now two new GUCs created by the Neon extension:
neon.allow_unstable_extensions: This is a session GUC which allows
a session to install and load unstable extensions.
neon.unstable_extensions: This is a comma-separated list of extension
names. We can check if a CREATE EXTENSION statement is attempting to
install an unstable extension, and if so, deny the request if
neon.allow_unstable_extensions is not set to true.
Signed-off-by: Tristan Partin <tristan@neon.tech>
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
Build the pgrag extensions (rag, rag_bge_small_en_v15, and
rag_jina_reranker_v1_tiny_en) as part of the compute node Dockerfile.
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
## Problem
Part of https://github.com/neondatabase/neon/issues/8623
## Summary of changes
Removed all aux-v1 config processing code. Note that we persisted it
into the index part file, so we cannot really remove the field from
index part. I also kept the config item within the tenant config, but we
will not read it any more.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
The `WalAcceptor` main loop currently uses two nested loops to consume
inbound messages. This makes it hard to slot in periodic events like
metrics collection. It also duplicates the event processing code, and assumes
all messages in steady state are AppendRequests (other messages types may
be dropped if following an AppendRequest).
## Summary of changes
Refactor the `WalAcceptor` loop to be event driven.
## Problem
Based on a recent proxy deployment issue, we deployed another proxy
version (proxy-scram), which was not needed when deploying a specific
proxy type. we have
[PR](https://github.com/neondatabase/infra/pull/2142) to update on the
infra branch and need to update CI in this repo which triggers proxy
deployment.
## Summary of changes
- Update proxy deployment flag
## Checklist before requesting a review
- [x] I have performed a self-review of my code.
- [ ] If it is a core feature, I have added thorough tests.
- [ ] Do we need to implement analytics? if so did you add the relevant
metrics to the dashboard?
- [ ] If this PR requires public announcement, mark it with
/release-notes label and add several sentences in this section.
## Checklist before merging
- [ ] Do not forget to reformat commit message to not include the above
checklist
virtio-serial is much more performant than /dev/console emulation,
therefore, is much more suitable for the verbose logs inside vm. This
commit changes routing for pgbouncer logs, since we've recently noticed
it can emit large volumes of logs.
Manually tested on staging by pinning a compute image to my test
project.
Should help with https://github.com/neondatabase/cloud/issues/19072
## Problem
We haven't historically taken this API route where we would onboard a
tenant to the controller in detached state. It worked, but we didn't
have test coverage.
## Summary of changes
- Add a test that onboards a tenant to the storage controller in
Detached mode, and checks that deleting it without attaching it works as
expected.
## Problem
If something goes wrong with a live migration, we currently only have
awkward ways to interrupt that:
- Restart the storage controller
- Ask it to do some other modification/migration on the shard, which we
don't really want.
## Summary of changes
- Add a new `/cancel` control API, and storcon_cli wrapper for it, which
fires the Reconciler's cancellation token. This is just for on-call use
and we do not expect it to be used by any other services.
## Problem
When we use pull_timeline API on an evicted timeline, it gets downloaded
to serve the snapshot API request. That means that to evacuate all the
timelines from a node, the node needs enough disk space to download
partial segments from all timelines, which may not be physically the
case.
Closes: #8833
## Summary of changes
- Add a "try" variant of acquiring a residence guard, that returns None
if the timeline is offloaded
- During snapshot API handler, take a different code path if the
timeline isn't resident, where we just read the checkpoint and don't try
to read any segments.
In complement to
https://github.com/neondatabase/tokio-epoll-uring/pull/56.
## Problem
We want to make tokio-epoll-uring slots waiters queue depth observable
via Prometheus.
## Summary of changes
- Add `pageserver_tokio_epoll_uring_slots_submission_queue_depth`
metrics as a `Histogram`.
- Each thread-local tokio-epoll-uring system is given a `LocalHistogram`
to observe the metrics.
- Keep a list of `Arc<ThreadLocalMetrics>` used on-demand to flush data
to the shared histogram.
- Extend `Collector::collect` to report
`pageserver_tokio_epoll_uring_slots_submission_queue_depth`.
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
This PR does two things:
1. Obtain a `TimelineCreateGuard` object in `unoffload_timeline`. This
prevents two unoffload tasks from racing with each other. While they
already obtain locks for `timelines` and `offloaded_timelines`, they
aren't sufficient, as we have already constructed an entire timeline at
that point. We shouldn't ever have two `Timeline` objects in the same
process at the same time.
2. don't allow timeline creations for timelines that have been
offloaded. Obviously they already exist, so we should not allow
creation. the previous logic only looked at the timelines list.
Part of #8088
## Problem
The storage components take an entire `SafekeeperConf` during
construction, but only actually use the `no_sync` field. This makes it
hard to understand the storage inputs (which fields do they actually
care about?), and is also inconvenient for tests and benchmarks that
need to set up a lot of unnecessary boilerplate.
## Summary of changes
* Don't take the entire config, but pass in the `no_sync` field
explicitly.
* Take the timeline dir instead of `ttid` as an input, since it's the
only thing it cares about.
* Fix a couple of tests to not leak tempdirs.
* Various minor tweaks.
## Problem
The Postgres version in `TimelinePersistentState::empty()` is incorrect:
the major version should be multiplied by 10000.
## Summary of changes
Multiply the version by 10000.
## Problem
We have some known N^2 behaviors when it comes to large relation counts,
due to the monolithic encoding and full rewrites of of RelDirectory each
time a relation is added. Ordinarily our backpressure mechanisms give
"slow but steady" performance when creating/dropping/truncating
relations. However, in the case of a transaction abort, it is possible
for a single WAL record to drop an unbounded number of relations. The
results in an unavailable compute, as when it sends one of these
records, it can stall the pageserver's ingest for many minutes, even
though the compute only sent a small amount of WAL.
Closes https://github.com/neondatabase/neon/issues/9505
## Summary of changes
- Rewrite relation-dropping code to do one read/modify/write cycle of
RelDirectory, instead of doing it separately for each relation in a
loop.
- Add a test for the bug scenario encountered:
`test_tx_abort_with_many_relations`
The test has ~40s runtime on my workstation. About 1 second of that is
the part where we wait for ingest to catch up after a rollback, the rest
is the slowness of creating and truncating a large number of relations.
---------
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
# Context
In the PGDATA import code
(https://github.com/neondatabase/neon/pull/9218) I add a third way to
create timelines, namely, by importing from a copy of a vanilla PGDATA
directory in object storage.
For idempotency, I'm using the PGDATA object storage location
specification, which is stored in the IndexPart for the entire lifespan
of the timeline. When loading the timeline from remote storage, that
value gets stored inside `struct Timeline` and timeline creation
compares the creation argument with that value to determine idempotency
of the request.
# Changes
This PR refactors the existing idempotency handling of Timeline
bootstrap and branching such that we simply compare the
`CreateTimelineIdempotency` struct, using the derive-generated
`PartialEq` implementation.
Also, by spelling idempotency out in the type names, I find it adds a
lot of clarity.
The pathway to idempotency via requester-provided idempotency key also
becomes very straight-forward, if we ever want to do this in the future.
# Refs
* platform context: https://github.com/neondatabase/neon/pull/9218
* product context: https://github.com/neondatabase/cloud/issues/17507
* stacks on top of https://github.com/neondatabase/neon/pull/9366
neon.c is getting crowded and the logical replication slot monitor is
a good candidate for reorganization. It is very self-contained, and
being in a separate file will make it that much easier to find.
Signed-off-by: Tristan Partin <tristan@neon.tech>
Previously it inserted ~150MiB of WAL while expecting page fetching to
work in 1s (wait_lsn_timeout=1s). It failed in CI in debug builds.
Instead, just directly wait for the wanted condition, i.e. needed
safekeepers are reported in pageserver timed out waiting for WAL error
message. Also set NEON_COMPUTE_TESTING_BASEBACKUP_RETRIES to 1 in this
test and neighbour one, it reduces execution time from 2.5m to ~10s.
## Problem
`local_fs` doesn't return file sizes, which I need in PGDATA import
(#9218)
## Solution
Include file sizes in the result.
I would have liked to add a unit test, and started doing that in
* https://github.com/neondatabase/neon/pull/9510
by extending the common object storage tests
(`libs/remote_storage/tests/common/tests.rs`) to check for sizes as
well.
But it turns out that localfs is not even covered by the common object
storage tests and upon closer inspection, it seems that this area needs
more attention.
=> punt the effort into https://github.com/neondatabase/neon/pull/9510
This PR adds a pageserver mgmt API to scan a layer file for disposable
keys.
It hooks it up to the sharding compaction test, demonstrating that we're
not filtering out all disposable keys.
This is extracted from PGDATA import
(https://github.com/neondatabase/neon/pull/9218)
where I do the filtering of layer files based on `is_key_disposable`.
Fixes#9098
## Problem
`test_readonly_node_gc` is flaky. As shown in [Allure
Report](https://neon-github-public-dev.s3.amazonaws.com/reports/pr-9469/11444519440/index.html#suites/3ccffb1d100105b98aed3dc19b717917/2c02073738fa2b39),
we would get a `AssertionError: No layers should be removed, old layers
are guarded by leases.` after the test restarts pageservers or after
reconfigure pageservers.
During the investigation, we found that the layers has LSN (`0/1563088`)
greater than the LSN (`0x1562000`) protected by the lease. For instance,
**Layers removed**
<pre>
000000067F00000005000034540100000000-000000067F00000005000040050100000000__000000000<b><i>1563088</i></b>-00000001
(shard 0002)
000000068000000000000017E20000000001-010000000100000001000000000000000001__000000000<b><i>1563088</i></b>-00000001
(shard 0002)
</pre>
**Lsn Lease Granted**
<pre>
handle_make_lsn_lease{lsn=<b><i>0/1562000</i></b> shard_id=0002
shard_id=0002}: lease created, valid until 2024-10-21
</pre>
This means that these layers are not guarded by the leases: they are in
"future", not visible to the static endpoint.
## Summary of changes
- Remove the assertion layers_removed == 0 after trigger timeline GC
while holding the lease. Instead rely on the successful execution of
the`SELECT` query to test lease validity.
- Improve test logging
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
Before, we didn't copy over the `index-part.json` of offloaded timelines
to the new shard's location, resulting in the new shard not knowing the
timeline even exists.
In #9444, we copy over the manifest, but we also need to do this for
`index-part.json`.
As the operations to do are mostly the same between offloaded and
non-offloaded timelines, we can iterate over all of them in the same
loop, after the introduction of a `TimelineOrOffloadedArcRef` type to
generalize over the two cases. This is analogous to the deletion code
added in #8907.
The added test also ensures that the sharded archival config endpoint
works, something that has not yet been ensured by tests.
Part of #8088
## Problem
https://github.com/neondatabase/neon/pull/9492 added a metric to track
the total count of block gaps filled on rel extend. More context is
needed to understand when this happens. The current theory is that it
may only happen on pg 14 and pg 15 since they do not WAL log relation extends.
## Summary of Changes
A rate limited log is added.
# Problem
Timeline creation can either be bootstrap or branch.
The distinction is made based on whether the `ancestor_*` fields are
present or not.
In the PGDATA import code
(https://github.com/neondatabase/neon/pull/9218), I add a third variant
to timeline creation.
# Solution
The above pushed me to refactor the code in Pageserver to distinguish
the different creation requests through enum variants.
There is no externally observable effect from this change.
On the implementation level, a notable change is that the acquisition of
the `TimelineCreationGuard` happens later than before. This is necessary
so that we have everything in place to construct the
`CreateTimelineIdempotency`. Notably, this moves the acquisition of the
creation guard _after_ the acquisition of the `gc_cs` lock in the case
of branching. This might appear as if we're at risk of holding `gc_cs`
longer than before this PR, but, even before this PR, we were holding
`gc_cs` until after the `wait_completion()` that makes the timeline
creation durable in S3 returns. I don't see any deadlock risk with
reversing the lock acquisition order.
As a drive-by change, I found that the `create_timeline()` function in
`neon_local` is unused, so I removed it.
# Refs
* platform context: https://github.com/neondatabase/neon/pull/9218
* product context: https://github.com/neondatabase/cloud/issues/17507
* next PR stacked atop this one:
https://github.com/neondatabase/neon/pull/9501
## Problem
WAL ingest couples decoding of special records with their handling
(updates to the storage engine mostly).
This is a roadblock for our plan to move WAL filtering (and implicitly
decoding) to safekeepers since they cannot
do writes to the storage engine.
## Summary of changes
This PR decouples the decoding of the special WAL records from their
application. The changes are done in place
and I've done my best to refrain from refactorings and attempted to
preserve the original code as much as possible.
Related: https://github.com/neondatabase/neon/issues/9335
Epic: https://github.com/neondatabase/neon/issues/9329
part of https://github.com/neondatabase/neon/issues/9114,
https://github.com/neondatabase/neon/issues/8836,
https://github.com/neondatabase/neon/issues/8362
The split layer writer code can be used in a more general way: the
caller puts unfinished writers into the batch layer writer and let batch
layer writer to ensure the atomicity of the layer produces.
## Summary of changes
* Add batch layer writer, which atomically finishes the layers.
`BatchLayerWriter::finish` is simply a copy-paste from previous split
layer writers.
* Refactor split writers to use the batch layer writer.
* The current split writer tests cover all code path of batch layer
writer.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We have `git config --global --add safe.directory ...` leftovers from the
past, but `actions/checkout` does it by default (since v3.0.2, we use v4)
## Summary of changes
- Remove `git config --global --add safe.directory ...` hack
## Problem
When a pageserver is misbehaving (e.g. we hit an ingest bug or something
is pathologically slow), the storage controller could get stuck in the
part of live migration that waits for LSNs to catch up. This is a
problem, because it can prevent us migrating the troublesome tenant to
another pageserver.
Closes: https://github.com/neondatabase/cloud/issues/19169
## Summary of changes
- Respect Reconciler::cancel during await_lsn.
A sizeof on a pointer on a 64 bit machine is 8 bytes whereas
Entry::old_name is a 64 byte array of characters. There was most likely
no fallout since the string would start with NUL bytes, but best to fix
nonetheless.
Signed-off-by: Tristan Partin <tristan@neon.tech>
## Problem
Filling the gap in with zeroes is annoying for sharded ingest. We are
not sure it even happens in reality.
## Summary of Changes
Add one global counter which tracks how many such gap blocks we filled
on relation extends. We can add more metrics once we understand the
scope.
## Problem
Occasionally, we get failures to start the storage controller's db with
errors like:
```
aborting due to panic at /__w/neon/neon/control_plane/src/background_process.rs:349:67:
claim pid file: lock file
Caused by:
file is already locked
```
e.g.
https://neon-github-public-dev.s3.amazonaws.com/reports/pr-9428/11380574562/index.html#/testresult/1c68d413ea9ecd4a
This is happening in a stop,start cycle during a test. Presumably the
pidfile from the startup background process is still held at the point
we stop, because we let pg_ctl keep running in the background.
## Summary of changes
- Refactor pg_ctl invocations into a helper
- In the controller's `start` function, use pg_ctl & a wait loop for
pg_isready, instead of using background_process
---------
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
We can't call pg_current_wal_lsn() if we are a standby instance (read
replica). Any attempt to call this function while in recovery results
in:
ERROR: recovery is in progress
Signed-off-by: Tristan Partin <tristan@neon.tech>
similar to https://github.com/neondatabase/neon/pull/8841, we make the
delta layer writer atomic when finishing the layers.
## Summary of changes
* `put_value` not taking discard fn anymore
* `finish` decides what layers to keep
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Persist timeline offloaded state to S3.
Right now, as of #8907, at each restart of the pageserver, all offloaded
state is lost, so we load the full timeline again. As it starts with an
empty local directory, we might potentially download some files again,
leading to downloads that are ultimately wasteful.
This patch adds support for persisting the offloaded state, allowing us
to never load offloaded timelines in the first place. The persistence
feature is facilitated via a new file in S3 that is tenant-global, which
contains a list of all offloaded timelines. It is updated each time we
offload or unoffload a timeline, and otherwise never touched.
This choice means that tenants where no offloading is happening will not
immediately get a manifest, keeping the change very minimal at the
start.
We leave generation support for future work. It is important to support
generations, as in the worst case, the manifest might be overwritten by
an older generation after a timeline has been unoffloaded (and
unarchived), so the next pageserver process instantiation might wrongly
believe that some timeline is still offloaded even though it should be
active.
Part of #9386, #8088
## Problem
If the environment variables `COMPATIBILITY_NEON_BIN` or
`COMPATIBILITY_POSTGRES_DISTRIB_DIR` are not set (this is usual during a
local run), the tests with the versions mix cannot run.
## Summary of changes
If these variables are not set turn off the version mix.
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
## Problem
Previously, figuring out how many tenant shards were managed by a
storage controller was typically done by peeking at the database or
calling into the API. A metric makes it easier to monitor, as
unexpectedly increasing shard counts can be indicative of problems
elsewhere in the system.
## Summary of changes
- Add metrics `storage_controller_pageserver_nodes` (updated on node
CRUD operations from Service) and `storage_controller_tenant_shards`
(updated RAII-style from TenantShard)
At least as far as removing individual files goes, this is the best
pattern for removing. I can't say the same for removing directories, but
I went ahead and changed those to `$(RM) -r` anyway.
Signed-off-by: Tristan Partin <tristan@neon.tech>
Always do timeline init through atomic rename of temp directory. Add
GlobalTimelines::load_temp_timeline which does this, and use it from
both pull_timeline and basic timeline creation. Fixes a collection
of issues:
- previously timeline creation didn't really flushed cfile to disk
due to 'nothing to do if state didn't change' check;
- even if it did, without tmp dir it is possible to lose the cfile
but leave timeline dir in place, making it look corrupted;
- tenant directory creation fsync was missing in timeline creation;
- pull_timeline is now protected from concurrent both itself and
timeline creation;
- now global timelines map entry got special CreationInProgress
entry type which prevents from anyone getting access to timeline
while it is being created (previously one could get access to it,
but it was locked during creation, which is valid but confusing if
creation failed).
fixes#8927
Add a way to list the offloaded timelines.
Before, one had to look at logs to figure out if a timeline has been
offloaded or not, or use the non-presence of a certain timeline in the
list of normal timelines. Now, one can list them directly.
Part of #8088
Part of #8130
## Problem
Pageserver previously goes through the kernel page cache for all the
IOs. The kernel page cache makes light-loaded pageserver have deceptive
fast performance. Using direct IO would offer predictable latencies of
our virtual file IO operations.
In particular for reads, the data pages also have an extremely low
temporal locality because the most frequently accessed pages are cached
on the compute side.
## Summary of changes
This PR enables pageserver to use direct IO for delta layer and image
layer reads. We can ship them separately because these layers are
write-once, read-many, so we will not be mixing buffered IO with direct
IO.
- implement `IoBufferMut`, an buffer type with aligned allocation
(currently set to 512).
- use `IoBufferMut` at all places we are doing reads on image + delta
layers.
- leverage Rust type system and use `IoBufAlignedMut` marker trait to
guarantee that the input buffers for the IO operations are aligned.
- page cache allocation is also made aligned.
_* in-memory layer reads and the write path will be shipped separately._
## Testing
Integration test suite run with O_DIRECT enabled:
https://github.com/neondatabase/neon/pull/9350
## Performance
We evaluated performance based on the `get-page-at-latest-lsn`
benchmark. The results demonstrate a decrease in the number of IOps, no
sigificant change in the latency mean, and an slight improvement on the
p99.9 and p99.99 latencies.
[Benchmark](https://www.notion.so/neondatabase/Benchmark-O_DIRECT-for-image-and-delta-layers-2024-10-01-112f189e00478092a195ea5a0137e706?pvs=4)
## Rollout
We will add `virtual_file_io_mode=direct` region by region to enable
direct IO on image + delta layers.
Signed-off-by: Yuchen Liang <yuchen@neon.tech>
Part of https://github.com/neondatabase/neon/issues/8836
## Summary of changes
This pull request makes the image layer split writer atomic when
finishing the layers. All the produced layers either finish at the same
time, or discard at the same time. Note that this does not prevent
atomicity when crash, but anyways, it will be cleaned up on pageserver
restart.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
```
+ /tmp/neon/pg_install/v16/bin/psql '***' -c 'SELECT version()'
/tmp/neon/pg_install/v16/bin/psql: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /tmp/neon/pg_install/v16/bin/psql)
/tmp/neon/pg_install/v16/bin/psql: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by /tmp/neon/pg_install/v16/bin/psql)
/tmp/neon/pg_install/v16/bin/psql: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.32' not found (required by /tmp/neon/pg_install/v16/lib/libpq.so.5)
/tmp/neon/pg_install/v16/bin/psql: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /tmp/neon/pg_install/v16/lib/libpq.so.5)
/tmp/neon/pg_install/v16/bin/psql: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by /tmp/neon/pg_install/v16/lib/libpq.so.5)
```
## Summary of changes
- Use `build-tools:pinned-bookworm` whenever we download Neon artefact
## Problem
Our dockerfiles, for some historical reason, have unconventional names
`Dockerfile.<something>`, and some tools (like GitHub UI) fail to highlight
the syntax in them.
> Some projects may need distinct Dockerfiles for specific purposes. A
common convention is to name these `<something>.Dockerfile`
From: https://docs.docker.com/build/concepts/dockerfile/#filename
## Summary of changes
- Rename `Dockerfile.build-tools` -> `build-tools.Dockerfile`
- Rename `compute/Dockerfile.compute-node` ->
`compute/compute-node.Dockerfile`
# `!failure() && !cancelled()` is required because the workflow transitively depends on the job that can be skipped: `push-to-acr-dev` and `push-to-acr-prod`
storage_broker={version="0.1",path="./storage_broker/"}# Note: main broker code is inside the binary crate, so linking with the library shouldn't be heavy.
@@ -21,8 +21,10 @@ The Neon storage engine consists of two major components:
See developer documentation in [SUMMARY.md](/docs/SUMMARY.md) for more information.
## Running local installation
## Running a local development environment
Neon can be run on a workstation for small experiments and to test code changes, by
following these instructions.
#### Installing dependencies on Linux
1. Install build dependencies and other applicable packages
@@ -132,7 +134,7 @@ make -j`sysctl -n hw.logicalcpu` -s
To run the `psql` client, install the `postgresql-client` package or modify `PATH` and `LD_LIBRARY_PATH` to include `pg_install/bin` and `pg_install/lib`, respectively.
To run the integration tests or Python scripts (not required to use the code), install
Python (3.9 or higher), and install the python3 packages using `./scripts/pysync` (requires [poetry>=1.8](https://python-poetry.org/)) in the project directory.
Python (3.11 or higher), and install the python3 packages using `./scripts/pysync` (requires [poetry>=1.8](https://python-poetry.org/)) in the project directory.
#### Running neon database
@@ -238,7 +240,7 @@ postgres=# select * from t;
> cargo neon stop
```
More advanced usages can be found at [Control Plane and Neon Local](./control_plane/README.md).
More advanced usages can be found at [Local Development Control Plane (`neon_local`))](./control_plane/README.md).
RUN curl -fsSL "https://github.com/protocolbuffers/protobuf/releases/download/v${PROTOC_VERSION}/protoc-${PROTOC_VERSION}-linux-$(uname -m | sed 's/aarch64/aarch_64/g').zip" -o "protoc.zip"\
@@ -116,33 +196,24 @@ RUN set -e \
# It includes several bug fixes on top on v2.0 release (https://github.com/linux-test-project/lcov/compare/v2.0...master)
# And patches from us:
# - Generates json file with code coverage summary (https://github.com/neondatabase/lcov/commit/426e7e7a22f669da54278e9b55e6d8caabd00af0.tar.gz)
@@ -54,24 +54,30 @@ ALTER ROLE regress_passwd2_new RENAME TO regress_passwd2;
@@ -54,24 +54,16 @@ ALTER ROLE regress_passwd2_new RENAME TO regress_passwd2;
-- passwords.
SET password_encryption = 'md5';
-- encrypt with MD5
-ALTER ROLE regress_passwd2 PASSWORD 'foo';
--- already encrypted, use as they are
-ALTER ROLE regress_passwd1 PASSWORD 'md5cd3578025fe2c3d7ed1b9a9b26238b70';
-ALTER ROLE regress_passwd3 PASSWORD 'SCRAM-SHA-256$4096:VLK4RMaQLCvNtQ==$6YtlR4t69SguDiwFvbVgVZtuz6gpJQQqUMZ7IQJK5yI=:ps75jrHeYU4lXCcXI4O8oIdJ3eO8o2jirjruw9phBTo=';
+ALTER ROLE regress_passwd2 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- already encrypted, use as they are
ALTER ROLE regress_passwd1 PASSWORD 'md5cd3578025fe2c3d7ed1b9a9b26238b70';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
ALTER ROLE regress_passwd3 PASSWORD 'SCRAM-SHA-256$4096:VLK4RMaQLCvNtQ==$6YtlR4t69SguDiwFvbVgVZtuz6gpJQQqUMZ7IQJK5yI=:ps75jrHeYU4lXCcXI4O8oIdJ3eO8o2jirjruw9phBTo=';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
SET password_encryption = 'scram-sha-256';
-- create SCRAM secret
-ALTER ROLE regress_passwd4 PASSWORD 'foo';
--- already encrypted with MD5, use as it is
-CREATE ROLE regress_passwd5 PASSWORD 'md5e73a4b11df52a6068f8b39f90be36023';
--- This looks like a valid SCRAM-SHA-256 secret, but it is not
--- so it should be hashed with SCRAM-SHA-256.
-CREATE ROLE regress_passwd6 PASSWORD 'SCRAM-SHA-256$1234';
--- These may look like valid MD5 secrets, but they are not, so they
--- should be hashed with SCRAM-SHA-256.
--- trailing garbage at the end
-CREATE ROLE regress_passwd7 PASSWORD 'md5012345678901234567890123456789zz';
--- invalid length
-CREATE ROLE regress_passwd8 PASSWORD 'md501234567890123456789012345678901zz';
+ALTER ROLE regress_passwd4 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- already encrypted with MD5, use as it is
CREATE ROLE regress_passwd5 PASSWORD 'md5e73a4b11df52a6068f8b39f90be36023';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
-- This looks like a valid SCRAM-SHA-256 secret, but it is not
-- so it should be hashed with SCRAM-SHA-256.
CREATE ROLE regress_passwd6 PASSWORD 'SCRAM-SHA-256$1234';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
-- These may look like valid MD5 secrets, but they are not, so they
-- should be hashed with SCRAM-SHA-256.
-- trailing garbage at the end
CREATE ROLE regress_passwd7 PASSWORD 'md5012345678901234567890123456789zz';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
-- invalid length
CREATE ROLE regress_passwd8 PASSWORD 'md501234567890123456789012345678901zz';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd5 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd6 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd7 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd8 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- Changing the SCRAM iteration count
SET scram_iterations = 1024;
CREATE ROLE regress_passwd9 PASSWORD 'alterediterationcount';
NOTICE: empty string is not a valid password, clearing password
@@ -1082,58 +1080,39 @@ index 8475231735..1afae5395f 100644
-(1 row)
+(0 rows)
-- Test with invalid stored and server keys.
--
-- The first is valid, to act as a control. The others have too long
-- stored/server keys. They will be re-hashed.
CREATE ROLE regress_passwd_sha_len0 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
CREATE ROLE regress_passwd_sha_len1 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96RqwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
CREATE ROLE regress_passwd_sha_len2 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=';
+ERROR: Received HTTP code 400 from control plane: {"error":"Neon only supports being given plaintext passwords"}
--- Test with invalid stored and server keys.
---
--- The first is valid, to act as a control. The others have too long
--- stored/server keys. They will be re-hashed.
-CREATE ROLE regress_passwd_sha_len0 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
-CREATE ROLE regress_passwd_sha_len1 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96RqwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
-CREATE ROLE regress_passwd_sha_len2 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=';
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd_sha_len0 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd_sha_len1 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd_sha_len2 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- Check that the invalid secrets were re-hashed. A re-hashed secret
-- should not contain the original salt.
SELECT rolname, rolpassword not like '%A6xHKoH/494E941doaPOYg==%' as is_rolpassword_rehashed
FROM pg_authid
WHERE rolname LIKE 'regress_passwd_sha_len%'
@@ -120,7 +109,7 @@ SELECT rolname, rolpassword not like '%A6xHKoH/494E941doaPOYg==%' as is_rolpassw
@@ -10,11 +10,11 @@ SET password_encryption = 'scram-sha-256'; -- ok
@@ -3200,25 +3192,61 @@ index 53e86b0b6c..f07cf1ec54 100644
-- check list of created entries
--
@@ -42,14 +42,14 @@ ALTER ROLE regress_passwd2_new RENAME TO regress_passwd2;
@@ -42,26 +42,18 @@ ALTER ROLE regress_passwd2_new RENAME TO regress_passwd2;
SET password_encryption = 'md5';
-- encrypt with MD5
-ALTER ROLE regress_passwd2 PASSWORD 'foo';
--- already encrypted, use as they are
-ALTER ROLE regress_passwd1 PASSWORD 'md5cd3578025fe2c3d7ed1b9a9b26238b70';
-ALTER ROLE regress_passwd3 PASSWORD 'SCRAM-SHA-256$4096:VLK4RMaQLCvNtQ==$6YtlR4t69SguDiwFvbVgVZtuz6gpJQQqUMZ7IQJK5yI=:ps75jrHeYU4lXCcXI4O8oIdJ3eO8o2jirjruw9phBTo=';
+ALTER ROLE regress_passwd2 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- already encrypted, use as they are
ALTER ROLE regress_passwd1 PASSWORD 'md5cd3578025fe2c3d7ed1b9a9b26238b70';
ALTER ROLE regress_passwd3 PASSWORD 'SCRAM-SHA-256$4096:VLK4RMaQLCvNtQ==$6YtlR4t69SguDiwFvbVgVZtuz6gpJQQqUMZ7IQJK5yI=:ps75jrHeYU4lXCcXI4O8oIdJ3eO8o2jirjruw9phBTo=';
SET password_encryption = 'scram-sha-256';
-- create SCRAM secret
-ALTER ROLE regress_passwd4 PASSWORD 'foo';
--- already encrypted with MD5, use as it is
-CREATE ROLE regress_passwd5 PASSWORD 'md5e73a4b11df52a6068f8b39f90be36023';
+ALTER ROLE regress_passwd4 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- already encrypted with MD5, use as it is
CREATE ROLE regress_passwd5 PASSWORD 'md5e73a4b11df52a6068f8b39f90be36023';
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd5 PASSWORD NEON_PASSWORD_PLACEHOLDER;
--- This looks like a valid SCRAM-SHA-256 secret, but it is not
--- so it should be hashed with SCRAM-SHA-256.
-CREATE ROLE regress_passwd6 PASSWORD 'SCRAM-SHA-256$1234';
--- These may look like valid MD5 secrets, but they are not, so they
--- should be hashed with SCRAM-SHA-256.
--- trailing garbage at the end
-CREATE ROLE regress_passwd7 PASSWORD 'md5012345678901234567890123456789zz';
--- invalid length
-CREATE ROLE regress_passwd8 PASSWORD 'md501234567890123456789012345678901zz';
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd6 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd7 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd8 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- Changing the SCRAM iteration count
SET scram_iterations = 1024;
@@ -78,13 +70,10 @@ ALTER ROLE regress_passwd_empty PASSWORD 'md585939a5ce845f1a1b620742e3c659e0a';
ALTER ROLE regress_passwd_empty PASSWORD 'SCRAM-SHA-256$4096:hpFyHTUsSWcR7O9P$LgZFIt6Oqdo27ZFKbZ2nV+vtnYM995pDh9ca6WSi120=:qVV5NeluNfUPkwm7Vqat25RjSPLkGeoZBQs6wVv+um4=';
SELECT rolpassword FROM pg_authid WHERE rolname='regress_passwd_empty';
--- Test with invalid stored and server keys.
---
--- The first is valid, to act as a control. The others have too long
--- stored/server keys. They will be re-hashed.
-CREATE ROLE regress_passwd_sha_len0 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
-CREATE ROLE regress_passwd_sha_len1 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96RqwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZI=';
-CREATE ROLE regress_passwd_sha_len2 PASSWORD 'SCRAM-SHA-256$4096:A6xHKoH/494E941doaPOYg==$Ky+A30sewHIH3VHQLRN9vYsuzlgNyGNKCh37dy96Rqw=:COPdlNiIkrsacU5QoxydEuOH6e/KfiipeETb/bPw8ZIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA=';
+-- Neon does not support encrypted passwords, use unencrypted instead
+CREATE ROLE regress_passwd_sha_len0 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd_sha_len1 PASSWORD NEON_PASSWORD_PLACEHOLDER;
+CREATE ROLE regress_passwd_sha_len2 PASSWORD NEON_PASSWORD_PLACEHOLDER;
-- Check that the invalid secrets were re-hashed. A re-hashed secret
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
