The LSN lease api actually accepts a tenant_shard_id, not a tenant_id.
But we put the Display of the tenant_shard_id into the tenant_id field.
This PR fixes it.
Refs
- fixes https://databricks.atlassian.net/browse/LKB-2930
## Problem
See [Slack
Channel](https://databricks.enterprise.slack.com/archives/C091LHU6NNB)
Dropping connection without resetting prefetch state can cause
request/response mismatch.
And lack of check response correctness in communicator_prefetch_lookupv
can cause data corruption.
## Summary of changes
1. Validate response before assignment to prefetch slot.
2. Consume prefetch requests before sending any other requests.
---------
Co-authored-by: Kosntantin Knizhnik <konstantin.knizhnik@databricks.com>
Co-authored-by: Konstantin Knizhnik <knizhnik@neon.tech>
The `--timelines-onto-safekeepers` flag is very consequential in the
sense that it controls every single timeline creation. However, we don't
have any automatic insight whether enabling the option will break things
or not.
The main way things can break is by misconfigured safekeepers, say they
are marked as paused in the storcon db. The best input so far we can
obtain via manually connecting via storcon_cli and listing safekeepers,
but this is cumbersome and manual so prone to human error.
So at storcon startup, do a simulated "test creation" in which we call
`timelines_onto_safekeepers` with the configuration provided to us, and
print whether it was successful or not. No actual timeline is created,
and nothing is written into the storcon db. The heartbeat info will not
have reached us at that point yet, but that's okay, because we still
fall back to safekeepers that don't have any heartbeat.
Also print some general scheduling policy stats on initial safekeeper
load.
Part of #11670.
## Problem
For the communicator, we need a rich Pageserver gRPC client.
Touches #11735.
Requires #12434.
## Summary of changes
This patch adds an initial rich Pageserver gRPC client. It supports:
* Sharded tenants across multiple Pageservers.
* Pooling of connections, clients, and streams for efficient resource
use.
* Concurrent use by many callers.
* Internal handling of GetPage bidirectional streams, with pipelining
and error handling.
* Automatic retries.
* Observability.
The client is still under development. In particular, it needs GetPage
batch splitting, shard map updates, and performance optimization. This
will be addressed in follow-up PRs.
The only differentiated handling of it is for `is_critical`, which in
turn is a `matches!()` on several variants of the `enum
CollectKeyspaceError`
which is the value contained insided
`CompactionError::CollectKeyspaceError`.
This PR introduces a new error for `repartition()`, allowing its
immediate
callers to inspect it like `is_critical` did.
A drive-by fix is more precise classification of WaitLsnError::BadState
when mapping to `tonic::Status`.
refs
- https://databricks.atlassian.net/browse/LKB-182
## Problem
close LKB-209
## Summary of changes
- We should not allow lease creation below the applied gc cutoff.
- Also removed the condition for `AttachedSingle`. We should always
check the lease against the gc cutoff in all attach modes.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We only trim the senders if we tried to send a message to them and
discovered that the channel is closed. This is problematic if the
pageserver keeps connecting while there's nothing to send back for the
shard. In this scenario we never trim down the senders list and can
panic due to the u8 limit.
## Summary of Changes
Trim down the dead senders before adding a new one.
Closes LKB-178
## Problem
We lost capability to explicitly disable the global eviction task (for
testing).
## Summary of changes
Add an `enabled` flag to `DiskUsageEvictionTaskConfig` to indicate
whether we should run the eviction job or not.
## Problem
The communicator will need gRPC channel/client/stream pools for
efficient reuse across many backends.
Touches #11735.
Requires #12396.
## Summary of changes
Adds three nested resource pools:
* `ChannelPool` for gRPC channels (i.e. TCP connections).
* `ClientPool` for gRPC clients (i.e. `page_api::Client`). Acquires
channels from `ChannelPool`.
* `StreamPool` for gRPC GetPage streams. Acquires clients from
`ClientPool`.
These are minimal functional implementations that will need further
improvements and performance optimization. However, the overall
structure is expected to be roughly final, so reviews should focus on
that.
The pools are not yet in use, but will form the foundation of a rich
gRPC Pageserver client used by the communicator (see #12462). This PR
also adds the initial crate scaffolding for that client.
See doc comments for details.
# TLDR
All changes are no-op except
1. publishing additional metrics.
2. problem VI
## Problem I
It has come to my attention that the Neon Storage Controller doesn't
correctly update its "observed" state of tenants previously associated
with PSs that has come back up after a local data loss. It would still
think that the old tenants are still attached to page servers and won't
ask more questions. The pageserver has enough information from the
reattach request/response to tell that something is wrong, but it
doesn't do anything about it either. We need to detect this situation in
production while I work on a fix.
(I think there is just some misunderstanding about how Neon manages
their pageserver deployments which got me confused about all the
invariants.)
## Summary of changes I
Added a `pageserver_local_data_loss_suspected` gauge metric that will be
set to 1 if we detect a problematic situation from the reattch response.
The problematic situation is when the PS doesn't have any local tenants
but received a reattach response containing tenants.
We can set up an alert using this metric. The alert should be raised
whenever this metric reports non-zero number.
Also added a HTTP PUT
`http://pageserver/hadron-internal/reset_alert_gauges` API on the
pageserver that can be used to reset the gauge and the alert once we
manually rectify the situation (by restarting the HCC).
## Problem II
Azure upload is 3x slower than AWS. -> 3x slower ingestion.
The reason for the slower upload is that Azure upload in page server is
much slower => higher flush latency => higher disk consistent LSN =>
higher back pressure.
## Summary of changes II
Use Azure put_block API to uploads a 1 GB layer file in 8 blocks in
parallel.
I set the put_block block size to be 128 MB by default in azure config.
To minimize neon changes, upload function passes the layer file path to
the azure upload code through the storage metadata. This allows the
azure put block to use FileChunkStreamRead to stream read from one
partition in the file instead of loading all file data in memory and
split it into 8 128 MB chunks.
## How is this tested? II
1. rust test_real_azure tests the put_block change.
3. I deployed the change in azure dev and saw flush latency reduces from
~30 seconds to 10 seconds.
4. I also did a bunch of stress test using sqlsmith and 100 GB TPCDS
runs.
## Problem III
Currently Neon limits the compaction tasks as 3/4 * CPU cores. This
limits the overall compaction throughput and it can easily cause
head-of-the-line blocking problems when a few large tenants are
compacting.
## Summary of changes III
This PR increases the limit of compaction tasks as `BG_TASKS_PER_THREAD`
(default 4) * CPU cores. Note that `CONCURRENT_BACKGROUND_TASKS` also
limits some other tasks `logical_size_calculation` and `layer eviction`
. But compaction should be the most frequent and time-consuming task.
## Summary of changes IV
This PR adds the following PageServer metrics:
1. `pageserver_disk_usage_based_eviction_evicted_bytes_total`: captures
the total amount of bytes evicted. It's more straightforward to see the
bytes directly instead of layers.
2. `pageserver_active_storage_operations_count`: captures the active
storage operation, e.g., flush, L0 compaction, image creation etc. It's
useful to visualize these active operations to get a better idea of what
PageServers are spending cycles on in the background.
## Summary of changes V
When investigating data corruptions, it's useful to search the base
image and all WAL records of a page up to an LSN, i.e., a breakdown of
GetPage@LSN request. This PR implements this functionality with two
tools:
1. Extended `pagectl` with a new command to search the layer files for a
given key up to a given LSN from the `index_part.json` file. The output
can be used to download the files from S3 and then search the file
contents using the second tool.
Example usage:
```
cargo run --bin pagectl index-part search --tenant-id 09b99ea3239bbb3b2d883a59f087659d --timeline-id 7bedf4a6995baff7c0421ff9aebbcdab --path ~/Downloads/corruption/index_part.json-0000000c-formatted --key 000000067F000080140000802100000D61BD --lsn 70C/BF3D61D8
```
Example output:
```
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F0000801400008028000002FEFF__000007089F0B5381-0000070C7679EEB9-0000000c
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000000000000000000000000000000000-000000067F0000801400008028000002F3F1__000006DD95B6F609-000006E2BA14C369-0000000c
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F000080140000802100001B0973__000006D33429F539-000006DD95B6F609-0000000c
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F00008014000080210000164D81__000006C6343B2D31-000006D33429F539-0000000b
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F0000801400008021000017687B__000006BA344FA7F1-000006C6343B2D31-0000000b
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F00008014000080210000165BAB__000006AD34613D19-000006BA344FA7F1-0000000b
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F0000801400000B180000000002-000000067F00008014000080210000137A39__0000069F34773461-000006AD34613D19-0000000b
tenants/09b99ea3239bbb3b2d883a59f087659d-0304/timelines/7bedf4a6995baff7c0421ff9aebbcdab/000000067F000080140000802100000D4000-000000067F000080140000802100000F0000__0000069F34773460-0000000b
```
2. Added a unit test to search the layer file contents. It's not
implemented part of `pagectl` because it depends on some test harness
code, which can only be used by unit tests.
Example usage:
```
cargo test --package pageserver --lib -- tenant::debug::test_search_key --exact --nocapture -- --tenant-id 09b99ea3239bbb3b2d883a59f087659d --timeline-id 7bedf4a6995baff7c0421ff9aebbcdab --data-dir /Users/chen.luo/Downloads/corruption --key 000000067F000080140000802100000D61BD --lsn 70C/BF3D61D8
```
Example output:
```
# omitted image for brievity
delta: 69F/769D8180: will_init: false, "OgAAALGkuwXwYp12nwYAAECGAAASIqLHAAAAAH8GAAAUgAAAIYAAAL1hDQD/DLGkuwUDAAAAEAAWAA=="
delta: 69F/769CB6D8: will_init: false, "PQAAALGkuwXotZx2nwYAABAJAAAFk7tpACAGAH8GAAAUgAAAIYAAAL1hDQD/CQUAEAASALExuwUBAAAAAA=="
```
## Problem VI
Currently when page service resolves shards from page numbers, it
doesn't fully support the case that the shard could be split in the
middle. This will lead to query failures during the tenant split for
either commit or abort cases (it's mostly for abort).
## Summary of changes VI
This PR adds retry logic in `Cache::get()` to deal with shard resolution
errors more gracefully. Specifically, it'll clear the cache and retry,
instead of failing the query immediately. It also reduces the internal
timeout to make retries faster.
The PR also fixes a very obvious bug in
`TenantManager::resolve_attached_shard` where the code tries to cache
the computed the shard number, but forgot to recompute when the shard
count is different.
---------
Co-authored-by: William Huang <william.huang@databricks.com>
Co-authored-by: Haoyu Huang <haoyu.huang@databricks.com>
Co-authored-by: Chen Luo <chen.luo@databricks.com>
Co-authored-by: Vlad Lazar <vlad.lazar@databricks.com>
Co-authored-by: Vlad Lazar <vlad@neon.tech>
This patch tightens up `page_api::Client`. It's mostly superficial
changes, but also adds a new constructor that takes an existing gRPC
channel, for use with the communicator connection pool.
## Problem
Some feature flags are used heavily on the critical path and we want the
"get feature flag" operation as cheap as possible.
## Summary of changes
Add a `test_remote_size_flag` as an example of such flags. In the
future, we can use macro to generate all those fields. The flag is
updated in the housekeeping loop. The retrieval of the flag is simply
reading an atomic flag.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
The goal of this code was to test out if resetting the broker
subscription helps alleviate the issues we've been seeing in staging.
Looks like it did the trick. However, the original version was too
eager.
## Summary of Changes
Only reset the stream when:
* we are waiting for WAL
* there's no connection candidates lined up
* we're not already connected to a safekeeper
The only call stack that can emit the `::AlreadyRunning` variant is
```
-> iteration_inner
-> iteration
-> compaction_iteration
-> compaction_loop
-> start_background_loops
```
And on that call stack, the only differentiated handling of it is its
invocations of
`log_compaction_error -> CompactionError::is_cancel()`, which returns
`true` for
`::AlreadyRunning`.
I think the condition of `AlreadyRunning` is severe; it really shouldn't
happen.
So, this PR starts treating it as something that is to be logged at
`ERROR` / `WARN`
level, depending on the `degrate_to_warning` argument to
`log_compaction_error`.
refs
- https://databricks.atlassian.net/browse/LKB-182
## Problem
Grafana Alloy in cluster mode seems to send duplicate "seconds" scrape
URL parameters
when one of its instances is disrupted.
## Summary of changes
Temporarily accept duplicate parameters as long as their value is
identical.
Looks can be deceiving: the match blocks in
`maybe_trip_compaction_breaker`
and at the end of `compact_with_options` seem like differentiated error
handling, but in reality, these branches are unreachable at runtime
because the only source of `CompactionError::Offload` within the
compaction code is at the end of `Tenant::compaction_iteration`.
We can simply map offload cancellation to CompactionError::Cancelled and
all other offload errors to ::Other, since there's no differentiated
handling for them in the compaction code.
Also, the OffloadError::RemoteStorage variant has no differentiated
handling, but was wrapping the remote storage anyhow::Error in a
`anyhow(thiserror(anyhow))` sandwich. This PR removes that variant,
mapping all RemoteStorage errors to `OffloadError::Other`.
Thereby, the sandwich is gone and we will get a proper anyhow backtrace
to the remote storage error location if when we debug-print the
OffloadError (or the CompactionError if we map it to that).
refs
- https://databricks.atlassian.net/browse/LKB-182
- the observation that there's no need for differentiated handling of
CompactionError::Offload was made in
https://databricks.slack.com/archives/C09254R641L/p1751286453930269?thread_ts=1751284317.955159&cid=C09254R641L
Before this PR, macOS builds would get clippy warning
```
warning: `tokio_epoll_uring::thread_local_system` does not refer to an existing function
```
The reason is that the `thread_local_system` function is only defined on
Linux.
Add `allow-invalid = true` to make macOS clippy pass, and manually test
that on Linux builds, clippy still fails when we use it.
refs
- https://databricks.slack.com/archives/C09254R641L/p1751917655527099
Co-authored-by: Christian Schwarz <Christian Schwarz>
## Problem
Deletion process does not calculate preferred nodes correctly - it
doesn't consider current tenant-shard layout among all pageservers.
## Summary of changes
- Added a schedule context calculation for node deletion
Co-authored-by: Aleksandr Sarantsev <aleksandr.sarantsev@databricks.com>
The introduction of the default lease deadline feature 9 months ago made
it so
that after PS restart, `.timeline_gc()` calls in Python tests are no-ops
for 10 minute after pageserver startup: the `gc_iteration()` bails with
`Skipping GC because lsn lease deadline is not reached`.
I did some impact analysis in the following PR. About 30 Python tests
are affected:
- https://github.com/neondatabase/neon/pull/12411
Rust tests that don't explicitly enable periodic GC or invoke GC
manually
are unaffected because we disable periodic GC by default in
the `TenantHarness`'s tenant config.
Two tests explicitly did `start_paused=true` + `tokio::time::advance()`,
but it would add cognitive and code bloat to each existing and future
test case that uses TenantHarness if we took that route.
So, this PR sets the default lease deadline feature in both Python
and Rust tests to zero by default. Tests that test the feature were
thus identified by failing the test:
- Python test `test_readonly_node_gc` + `test_lsn_lease_size`
- Rust test `test_lsn_lease`.
To accomplish the above, I changed the code that computes the initial
lease
deadline to respect the pageserver.toml's default tenant config, which
it didn't before (and I would consider a bug). The Python test harness
and the Rust TenantHarness test harness then simply set the default
tenant
config field to zero.
Drive-by:
- `test_lsn_lease_size` was writing a lot of data unnecessarily; reduce
the amount and speed up the test
refs
- PR that introduced default lease deadline:
https://github.com/neondatabase/neon/pull/9055/files
- fixes https://databricks.atlassian.net/browse/LKB-92
---------
Co-authored-by: Christian Schwarz <Christian Schwarz>
## Problem
Extension tests were previously run sequentially, resulting in
unnecessary wait time and underutilization of available CPU cores.
## Summary of changes
Tests are now executed in a customizable number of parallel threads
using separate database branches. This reduces overall test time by
approximately 50% (e.g., on my laptop, parallel test lasts 173s, while
sequential one lasts 340s) and increases the load on the pageserver,
providing better test coverage.
---------
Co-authored-by: Alexander Bayandin <alexander@neon.tech>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Co-authored-by: Alexey Masterov <alexey.masterov@databricks.com>
## Problem
Pageserver now writes errors in the log during the safekeeper migration.
Some errors are added to allowed errors, but "timeline not found in
global map" is not.
- Will be properly fixed in
https://github.com/neondatabase/neon/issues/12191
## Summary of changes
Add "timeline not found in global map" error in a list of allowed errors
in `test_safekeeper_migration_simple`
## Problem
Test is flaky due to the following warning in the logs:
```
Keeping extra secondaries: can't determine which of [NodeId(1), NodeId(2)] to remove (some nodes offline?)
```
Some nodes being offline is expected behavior in this test.
## Summary of changes
- Added `Keeping extra secondaries` to the list of allowed errors
- Improved logging for better debugging experience
Co-authored-by: Aleksandr Sarantsev <aleksandr.sarantsev@databricks.com>
## Problem
The endpoint filter cache is still unused because it's not yet reliable
enough to be used. It only consumes a lot of memory.
## Summary of changes
Remove the code. Needs a new design.
neondatabase/cloud#30634
See #11992 and #11961 for some examples of usecases.
This introduces a JSON serialization lib, designed for more flexibility
than serde_json offers.
## Dynamic construction
Sometimes you have dynamic values you want to serialize, that are not
already in a serde-aware model like a struct or a Vec etc. To achieve
this with serde, you need to implement a lot of different traits on a
lot of different new-types. Because of this, it's often easier to
give-in and pull all the data into a serde-aware model
(serde_json::Value or some intermediate struct), but that is often not
very efficient.
This crate allows full control over the JSON encoding without needing to
implement any extra traits. Just call the relevant functions, and it
will guarantee a correctly encoded JSON value.
## Async construction
Similar to the above, sometimes the values arrive asynchronously. Often
collecting those values in memory is more expensive than writing them as
JSON, since the overheads of `Vec` and `String` is much higher, however
there are exceptions.
Serializing to JSON all in one go is also more CPU intensive and can
cause lag spikes, whereas serializing values incrementally spreads out
the CPU load and reduces lag.
