Previously, we would wait for the LSN to be visible on whichever
timeline we happened to load at the start of the connection, then
proceed to look up the correct timeline for the key and do the read.
If the timeline holding the key was behind the timeline we used
for the LSN wait, then we might serve an apparently-successful read result
that actually contains data from behind the requested lsn.
If there is some secondary shard for a tenant on the same
node as an attached shard, the secondary shard could trip up
this code and cause page_service to incorrectly
get an error instead of finding the attached shard.
It has caveats such as creating half empty segment which can't be
offloaded. Instead we'll pursue approach of pull_timeline, seeding new state
from some peer.
## Problem
We need to add one more patch to pgbouncer (for
https://github.com/neondatabase/neon/issues/5801). I've decided to
cherry-pick all required patches to a pgbouncer fork
(`neondatabase/pgbouncer`) and use it instead.
See
https://github.com/neondatabase/pgbouncer/releases/tag/pgbouncer_1_21_0-neon-1
## Summary of changes
- Revert the previous patch (for deallocate/discard all) — the fork
already contains it.
- Remove `libssl-dev` dependency — we build pgbouncer without `openssl`
support.
- Clone git tag and build pgbouncer from source code.
## Problem
For context, this problem was observed in a research project where we
try to make neon run in multiple regions and I was asked by @hlinnaka to
make this PR.
In our project, we use the pageserver in a non-conventional way such
that we would send a larger number of requests to the pageserver than
normal (imagine postgres without the buffer pool). I measured the time
from the moment a WAL record left the safekeeper to when it reached the
pageserver
([code](e593db1f5a/pageserver/src/tenant/timeline/walreceiver/walreceiver_connection.rs (L282-L287)))
and observed that when the number of get_page_at_lsn requests was high,
the wal receiving time increased significantly (see the left side of the
graphs below).
Upon further investigation, I found that the delay was caused by this
line
d2ca410919/pageserver/src/tenant/timeline.rs (L2348)
The `get_layer_for_write` method is called for every value during WAL
ingestion and it tries to acquire layers write lock every time, thus
this results in high contention when read lock is acquired more
frequently.
## Summary of changes
It is unnecessary to call `get_layer_for_write` repeatedly for all
values in a WAL message since they would end up in the same memory layer
anyway, so I created the batched versions of `InMemoryLayer::put_value`,
`InMemoryLayer ::put_tombstone`, `Timeline::put_value`, and
`Timeline::put_tombstone`, that acquire the locks once for a batch of
values.
Additionally, `DatadirModification` is changed to store multiple
versions of uncommitted values, and `WalIngest::ingest_record()` can now
ingest records without immediately committing them.
With these new APIs, the new ingestion loop can be changed to commit for
every `ingest_batch_size` records. The `ingest_batch_size` variable is
exposed as a config. If it is set to 1 then we get the same behavior
before this change. I found that setting this value to 100 seems to work
the best, and you can see its effect on the right side of the above
graphs.
---------
Co-authored-by: John Spray <john@neon.tech>
## Problem
The version of pytest we were using emits a number of
DeprecationWarnings on latest python: these are fixed in latest release.
boto3 and python-dateutil also have deprecation warnings, but
unfortunately these aren't fixed upstream yet.
## Summary of changes
- Update pytest
- Update boto3 (this doesn't fix deprecation warnings, but by the time I
figured that out I had already done the update, and it's good hygiene
anyway)
The tool still needs a lot of work. These are the easiest fix and
feature:
- use similar adaptive config with s3 as remote_storage, use retries
- process only particular tenants
Tenants need to be from the correct region, they are not deduplicated,
but the feature is useful for re-checking small amount of tenants after
a large run.
To exercise MAX_SEND_SIZE sending from safekeeper; we've had a bug with WAL
records torn across several XLogData messages. Add failpoint to safekeeper to
slow down sending. Also check for corrupted WAL complains in standby log.
Make the test a bit simpler in passing, e.g. we don't need explicit commits as
autocommit is enabled by default.
https://neondb.slack.com/archives/C05L7D1JAUS/p1703774799114719https://github.com/neondatabase/cloud/issues/9057
Otherwise they are left orphaned when compute_ctl is terminated with a
signal. It was invisible most of the time because normally neon_local or k8s
kills postgres directly and then compute_ctl finishes gracefully. However, in
some tests compute_ctl gets stuck waiting for sync-safekeepers which
intentionally never ends because safekeepers are offline, and we want to stop
compute_ctl without leaving orphanes behind.
This is a quite rough approach which doesn't wait for children termination. A
better way would be to convert compute_ctl to async which would make waiting
easy.
## Problem
`cargo test` doesn't support timeouts
or junit output format
## Summary of changes
- Add `nextest` to `build-tools` image
- Switch `cargo test` with `cargo nextest` on CI
- Set timeout
## Problem
control plane should be able to return domain names and not just IP
addresses.
## Summary of changes
1. add regression tests
2. use rsplit to split the port from the back, then trim the ipv6
brackets
## Problem
Extension pg_stat_statements has function pg_stat_statements_reset().
In vanilla Postgres this function can only be called by superuser role
or other users/roles explicitly granted.
In Neon no end user can use superuser role.
Instead we have neon_superuser role.
We need to grant execute on pg_stat_statements_reset() to neon_superuser
## Summary of changes
Modify the Postgres v14, v15, v16 contrib in our compute docker file to
grant execute on pg_stat_statements_reset() to neon_superuser.
(Modifying it in our docker file is preferable to changes in
neondatabase/postgres because we want to limit the changes in our fork
that we have to carry with each new version of Postgres).
Note that the interface of proc/function pg_stat_statements_reset
changed in pg_stat_statements version 1.7
So for versions up to and including 1.6 we must
`GRANT EXECUTE ON FUNCTION pg_stat_statements_reset() TO
neon_superuser;`
and for versions starting from 1.7 we must
`GRANT EXECUTE ON FUNCTION pg_stat_statements_reset(Oid, Oid, bigint) TO
neon_superuser;`
If we just use `GRANT EXECUTE ON FUNCTION pg_stat_statements_reset() TO
neon_superuser;` for all version this results in the following error for
versions 1.7+:
```sql
neondb=> create extension pg_stat_statements;
ERROR: function pg_stat_statements_reset() does not exist
```
## 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?
- [x ] 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
## I have run the following test and could now invoke
pg_stat_statements_reset() using default user
```bash
(neon) peterbendel@Peters-MBP neon % kubectl get pods | grep compute-quiet-mud-88416983
compute-quiet-mud-88416983-74f4bf67db-crl4c 3/3 Running 0 7m26s
(neon) peterbendel@Peters-MBP neon % kubectl set image deploy/compute-quiet-mud-88416983 compute-node=neondatabase/compute-node-v15:7307610371
deployment.apps/compute-quiet-mud-88416983 image updated
(neon) peterbendel@Peters-MBP neon % psql postgresql://peterbendel:<secret>@ep-bitter-sunset-73589702.us-east-2.aws.neon.build/neondb
psql (16.1, server 15.5)
SSL connection (protocol: TLSv1.3, cipher: TLS_AES_256_GCM_SHA384, compression: off)
Type "help" for help.
neondb=> select version();
version
---------------------------------------------------------------------------------------------------
PostgreSQL 15.5 on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
(1 row)
neondb=> create extension pg_stat_statements;
CREATE EXTENSION
neondb=> select pg_stat_statements_reset();
pg_stat_statements_reset
--------------------------
(1 row)
```
- add pgbouncer_settings section to compute spec;
- add pgbouncer-connstr option to compute_ctl.
- add pgbouncer-ini-path option to compute_ctl. Default: /etc/pgbouncer/pgbouncer.ini
Apply pgbouncer config on compute start and respec to override default spec.
Save pgbouncer config updates to pgbouncer.ini to preserve them across pgbouncer restarts.
Remove confirm_wal_streamed; we already apply both write and flush positions of
the slot to commit_lsn which is fine because 1) we need to wake up waiters 2)
committed WAL can be fetched from safekeepers by neon_walreader now.
wp -> sk communication now uses neon_walreader which will fetch missing WAL on
demand from safekeepers, so doesn't need this anymore. Also, cap WAL download by
max_slot_wal_keep_size to be able to start compute if lag is too high.
It is similar to XLogReader, but when either requested segment is missing
locally or requested LSN is before basebackup_lsn NeonWALReader asynchronously
fetches WAL from one of safekeepers.
Patch includes walproposer switch to NeonWALReader, splitting wouldn't make much
sense as it is hard to test otherwise. This finally removes risk of pg_wal
explosion (as well as slow start time) when one safekeeper is lagging, at the
same time allowing to recover it.
In the future reader should also be used by logical walsender for similar
reasons (currently we download the tail on compute start synchronously).
The main test is test_lagging_sk. However, I also run it manually a lot varying
MAX_SEND_SIZE on both sides (on safekeeper and on walproposer), testing various
fragmentations (one side having small buffer, another, both), which brought up
https://github.com/neondatabase/neon/issues/6055
closes https://github.com/neondatabase/neon/issues/1012
## Problem
See https://neondb.slack.com/archives/C026T7K2YP9/p1702813041997959
## Summary of changes
Do not take in account invalidated slots when calculate restart_lsn
position for basebackup at 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>
Store the content of the `last-modified` and `etag` HTTP headers in
`Download`.
This serves both as the first step towards #6199 and as a preparation
for tests in #6155 .
## Problem
During startup, a tenant delete request might have to retry for many
minutes waiting for a tenant to enter Active state.
## Summary of changes
- Refactor delete_tenant into TenantManager: this is not a functional
change, but will avoid merge conflicts with
https://github.com/neondatabase/neon/pull/6105 later
- Add 412 responses to the swagger definition of this endpoint.
- Use Tenant::wait_to_become_active in `TenantManager::delete_tenant`
---------
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
## Problem
Fix build-tools image tag calculation for PRs.
Broken in https://github.com/neondatabase/neon/pull/6195
## Summary of changes
- Use `pinned` tag instead of `$GITHUB_RUN_ID` if there's no changes in
the dockerfile (and we don't build such image)
## Fixing GitHub workflow issue related to build and push images
## Summary of changes
Followup of PR#608[move docker file from build repo to neon to solve
issue some issues
The build started failing because it missed a validation in logic that
determines changes in the docker file
Also, all the dependent jobs were skipped because of the build and push
of the image job.
To address the above issue following changes were made
- we are adding validation to generate image tag even if it's a merge to
repo.
- All the dependent jobs won't skip even if the build and push image job
is skipped.
- We have moved the logic to generate a tag in the sub-workflow. As the
tag name was necessary to be passed to the sub-workflow it made sense to
abstract that away where it was needed and then store it as an output
variable so that downward dependent jobs could access the value.
- This made the dependency logic easy and we don't need complex
expressions to check the condition on which it will run
- An earlier PR was closed that tried solving a similar problem that has
some feedback and context before creating this PR
https://github.com/neondatabase/neon/pull/6175
## Checklist before requesting a review
- [x] Move the tag generation logic from the main workflow to the
sub-workflow of build and push the image
- [x] Add a condition to generate an image tag for a non-PR-related run
- [x] remove complex if the condition from the job if conditions
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
Co-authored-by: Abhijeet Patil <abhijeet@neon.tech>
## Problem
This is a comment only change.
To ensure that our benchmarking results are fair we need to have correct
stats in catalog. Otherwise optimizer chooses seq scan instead of index
only scan for some queries. Added comment to run vacuum after data prep.
This PR adds a component-level benchmarking utility for pageserver.
Its name is `pagebench`.
The problem solved by `pagebench` is that we want to put Pageserver
under high load.
This isn't easily achieved with `pgbench` because it needs to go through
a compute, which has signficant performance overhead compared to
accessing Pageserver directly.
Further, compute has its own performance optimizations (most
importantly: caches). Instead of designing a compute-facing workload
that defeats those internal optimizations, `pagebench` simply bypasses
them by accessing pageserver directly.
Supported benchmarks:
* getpage@latest_lsn
* basebackup
* triggering logical size calculation
This code has no automated users yet.
A performance regression test for getpage@latest_lsn will be added in a
later PR.
part of https://github.com/neondatabase/neon/issues/5771
## Problem
HTTP connection pool will grow without being pruned
## Summary of changes
Remove connection clients from pools once idle, or once they exit.
Periodically clear pool shards.
GC Logic:
Each shard contains a hashmap of `Arc<EndpointPool>`s.
Each connection stores a `Weak<EndpointPool>`.
During a GC sweep, we take a random shard write lock, and check that if
any of the `Arc<EndpointPool>`s are unique (using `Arc::get_mut`).
- If they are unique, then we check that the endpoint-pool is empty, and
sweep if it is.
- If they are not unique, then the endpoint-pool is in active use and we
don't sweep.
- Idle connections will self-clear from the endpoint-pool after 5
minutes.
Technically, the uniqueness of the endpoint-pool should be enough to
consider it empty, but the connection count check is done for
completeness sake.
