- avoid panic by checking for Ok(None) response from
tonic::Streaming::message() instead of just using unwrap()
- There was a race condition, if the caller sent the message, but the
receiver task concurrently received Ok(None) indicating the stream
was closed. (I didn't see that in action, but I think it could happen
by reading the code)
After reading the code a few times, I didn't quite understand what it
was, to be honest, or how it was going to be used. Remove it now to
reduce noise, but we can resurrect it from git history if we need it
in the future.
Add a new 'pageserver_connection_info' field in the compute spec. It
replaces the old 'pageserver_connstring' field with a more complicated
struct that includes both libpq and grpc URLs, for each shard (or only
one of the the URLs, depending on the configuration). It also includes
a flag suggesting which one to use; compute_ctl now uses it to decide
which protocol to use for the basebackup.
This is compatible with everything that's in production, because the
control plane never used the 'pageserver_connstring' field. That was
added a long time ago with the idea that it would replace the code
that digs the 'neon.pageserver_connstring' GUC from the list of
Postgres settings, but we never got around to do that in the control
plane. Hence, it was only used with neon_local. But the plan now is to
pass the 'pageserver_connection_info' from the control plane, and once
that's fully deployed everywhere, the code to parse
'neon.pageserver_connstring' in compute_ctl can be removed.
The 'grpc' flag on an endpoint in endpoint config is now more of a
suggestion. Compute_ctl gets both URLs, so it can choose to use libpq
or grpc as it wishes. It currently always obeys the 'prefer_grpc' flag
that's part of the connection info though. Postgres however uses grpc
iff the new rust-based communicator is enabled.
TODO/plan for the control plane:
- Start to pass `pageserver_connection_info` in the spec file.
- Also keep the current `neon.pageserver_connstring` setting for now,
for backwards compatibility with old computes
After that, the `pageserver_connection_info.prefer_grpc` flag in the
spec file can be used to control whether compute_ctl uses grpc or
libpq. The actual compute's grpc usage will be controlled by the
`neon.enable_new_communicator` GUC. It can be set separately from
'prefer_grpc'.
Later:
- Once all old computes are gone, remove the code to pass
`neon.pageserver_connstring`
## Problem
Part of #11813
## Summary of changes
The current interval is 30s and it costs a lot of $$$. This patch
reduced it to 600s refresh interval (which means that it takes 10min for
feature flags to propagate from UI to the pageserver). In the future we
can let storcon retrieve the feature flags and push it to pageservers.
We can consider creating a new release or we can postpone this to the
week after the next week.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Previously, we couldn't read from an in-memory layer while a batch was
being written to it. Vice-versa, we couldn't write to it while there
was an on-going read.
## Summary of Changes
The goal of this change is to improve concurrency. Writes happened
through a &mut self method so the enforcement was at the type system
level.
We attempt to improve by:
1. Adding interior mutability to EphemeralLayer. This involves wrapping
the buffered writer in a read-write lock.
2. Minimise the time that the read lock is held for. Only hold the read
lock while reading from the buffers (recently flushed or pending
flush). If we need to read from the file, drop the lock and allow IO
to be concurrent.
The new benchmark variants with concurrent reads improve between 70 to
200 percent (against main).
Benchmark results are in this
[commit](891f094ce6).
## Future Changes
We can push the interior mutability into the buffered writer. The
mutable tail goes under a read lock, the flushed part goes into an
ArcSwap and then we can read from anything that is flushed _without_ any
locking.
gRPC base backups send a stream of fixed-size 64KB chunks.
pagebench basebackup with compression enabled shows this to reduce
throughput:
* 64 KB: 55 RPS
* 128 KB: 69 RPS
* 256 KB: 73 RPS
* 1024 KB: 73 RPS
This patch sets the base backup chunk size to 256 KB.
## Problem
In #12217, we began passing the stripe size in reattach responses, and
persisting it in the on-disk state. This is necessary to ensure the
storage controller and Pageserver have a consistent view of the intended
stripe size of unsharded tenants, which will be used for splits that do
not specify a stripe size. However, for backwards compatibility, these
stripe sizes were optional.
## Summary of changes
Make the stripe sizes required for reattach responses and on-disk
location configs. These will always be provided by the previous
(current) release.
## Problem
Timeline GC is very aggressive with regards to layer map locking.
We've seen timelines with loads of layers in production that hold the
write lock for the layer map for 30 minutes at a time.
This blocks reads and the write path to some extent.
## Summary of changes
Determining the set of layers to GC is done under the read lock.
Applying the updates is done under the write lock.
Previously, everything was done under write lock.
This improves `pagebench getpage-latest-lsn` gRPC support by:
* Using `page_api::Client`.
* Removing `--protocol`, and using the `page-server-connstring` scheme
instead.
* Adding `--compression` to enable zstd compression.
## Problem
Pagebench does not support gRPC for `basebackup` benchmarks.
Requires #12243.
Touches #11728.
## Summary of changes
Add gRPC support via gRPC connstrings, e.g. `pagebench basebackup
--page-service-connstring grpc://localhost:51051`.
Also change `--gzip-probability` to `--no-compression`, since this must
be specified per-client for gRPC.
Introduce a separate `postgres_ffi_types` crate which contains a few
types and functions that were used in the API. `postgres_ffi_types` is a
much small crate than `postgres_ffi`, and it doesn't depend on bindgen
or the Postgres C headers.
Move NeonWalRecord and Value types to wal_decoder crate. They are only
used in the pageserver-safekeeper "ingest" API. The rest of the ingest
API types are defined in wal_decoder, so move these there as well.
## Problem
The gRPC page service should support compression.
Requires #12111.
Touches #11728.
Touches https://github.com/neondatabase/cloud/issues/25679.
## Summary of changes
Add support for gzip and zstd compression in the server, and a client
parameter to enable compression.
This will need further benchmarking under realistic network conditions.
## Problem
GC info is an input to updating layer visibility.
Currently, gc info is updated on timeline activation and visibility is
computed on tenant attach, so we ignore branch points and compute
visibility by taking all layers into account.
Side note: gc info is also updated when timelines are created and
dropped. That doesn't help because we create the timelines in
topological order from the root. Hence the root timeline goes first,
without context of where the branch points are.
The impact of this in prod is that shards need to rehydrate layers after
live migration since the non-visible ones were excluded from the
heatmap.
## Summary of Changes
Move the visibility calculation into tenant attachment instead of
activation.
## Problem
Full basebackups are used in tests, and may be useful for debugging as
well, so we should support them in the gRPC API.
Touches #11728.
## Summary of changes
Add `GetBaseBackupRequest::full` to generate full base backups.
The libpq implementation also allows specifying `prev_lsn` for full
backups, i.e. the end LSN of the previous WAL record. This is omitted in
the gRPC API, since it's not used by any tests, and presumably of
limited value since it's autodetected. We can add it later if we find
that we need it.
## Problem
When converting `proto::GetPageRequest` into `page_api::GetPageRequest`
and validating the request, errors are returned as `tonic::Status`. This
will tear down the GetPage stream, which is disruptive and unnecessary.
## Summary of changes
Emit invalid request errors as `GetPageResponse` with an appropriate
`status_code` instead.
Also move the conversion from `tonic::Status` to `GetPageResponse` out
into the stream handler.
## Problem
The cache was introduced as a hackathon project and the only supported
limit was the number of entries.
The basebackup entry size may vary. We need to have more control over
disk space usage to ship it to production.
- Part of https://github.com/neondatabase/cloud/issues/29353
## Summary of changes
- Store the size of entries in the cache and use it to limit
`max_total_size_bytes`
- Add the size of the cache in bytes to metrics.
## Problem
Pageservers now expose a gRPC API on a separate address and port. This
must be registered with the storage controller such that it can be
plumbed through to the compute via cplane.
Touches #11926.
## Summary of changes
This patch registers the gRPC address and port with the storage
controller:
* Add gRPC address to `nodes` database table and `NodePersistence`, with
a Diesel migration.
* Add gRPC address in `NodeMetadata`, `NodeRegisterRequest`,
`NodeDescribeResponse`, and `TenantLocateResponseShard`.
* Add gRPC address flags to `storcon_cli node-register`.
These changes are backwards-compatible, since all structs will ignore
unknown fields during deserialization.
## Problem
The gRPC base backup implementation has a few issues: chunks are not
properly bounded, and it's not possible to omit the LSN.
Touches #11728.
## Summary of changes
* Properly bound chunks by using a limited writer.
* Use an `Option<Lsn>` rather than a `ReadLsn` (the latter requires an
LSN).
## Problem
In our environment, we don't always have access to the pagectl tool on
the pageserver. We have to download the page files to local env to
introspect them. Hence, it'll be useful to be able to parse the local
files using `pagectl`.
## Summary of changes
* Add `dump-layer-local` to `pagectl` that takes a local path as
argument and returns the layer content:
```
cargo run -p pagectl layer dump-layer-local ~/Desktop/000000067F000040490002800000FFFFFFFF-030000000000000000000000000000000002__00003E7A53EDE611-00003E7AF27BFD19-v1-00000001
```
* Bonus: Fix a bug in `pageserver/ctl/src/draw_timeline_dir.rs` in which
we don't filter out temporary files.
## Problem
The `page_api` domain types are missing a few derives.
## Summary of changes
Add `Clone`, `Copy`, and `Debug` derives for all types where
appropriate.
## Problem
We would easily hit this limit for a tenant running for enough long
time.
## Summary of changes
Remove the max key limit for time-travel recovery if the command is
running locally.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
* Inside `compact_with_gc_inner`, there is a similar log line:
db24ba95d1/pageserver/src/tenant/timeline/compaction.rs (L3181-L3187)
* Also, I think it would be useful when debugging to have the ability to
select a particular sub-compaction job (e.g., `1/100`) to see all the
logs for that job.
## Summary of changes
* Attach a span to the `compact_with_gc_inner`.
CC: @skyzh
## Problem
The location config (which includes the stripe size) is stored on
pageserver disk.
For unsharded tenants we [do not include the shard identity in the
serialized
description](ad88ec9257/pageserver/src/tenant/config.rs (L64-L66)).
When the pageserver restarts, it reads that configuration and will use
the stripe size from there
and rely on storcon input from reattach for generation and mode.
The default deserialization is ShardIdentity::unsharded. This has the
new default stripe size of 2048.
Hence, for unsharded tenants we can be running with a stripe size
different from that the one in the
storcon observed state. This is not a problem until we shard split
without specifying a stripe size (i.e. manual splits via the UI or
storcon_cli). When that happens the new shards will use the 2048 stripe
size until storcon realises and switches them back. At that point it's
too late, since we've ingested data with the wrong stripe sizes.
## Summary of changes
Ideally, we would always have the full shard identity on disk. To
achieve this over two releases we do:
1. Always persist the shard identity in the location config on the PS.
2. Storage controller includes the stripe size to use in the re attach
response.
After the first release, we will start persisting correct stripe sizes
for any tenant shard that the storage controller
explicitly sends a location_conf. After the second release, the
re-attach change kicks in and we'll persist the
shard identity for all shards.
## Problem
We hold the layer map for too long on occasion.
## Summary of changes
This should help us identify the places where it's happening from.
Related https://github.com/neondatabase/neon/issues/12182
## Problem
There is this TODO in code:
https://github.com/neondatabase/neon/blob/main/pageserver/src/tenant/mgr.rs#L300-L302
This is an old TODO by @jcsp.
## Summary of changes
This PR addresses the TODO. Specifically, it removes a global static
`TENANTS`. Instead the `TenantManager` now directly manages the tenant
map. Enhancing abstraction.
Essentially, this PR moves all module-level methods to inside the
implementation of `TenantManager`.
## Problem
part of https://github.com/neondatabase/neon/issues/11813
## Summary of changes
Collect pageserver hostname property so that we can use it in the
PostHog UI. Not sure if this is the best way to do that -- open to
suggestions.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
We are getting tenants with a lot of branches and num of timelines is a
good indicator of pageserver loads. I added this metrics to help us
better plan pageserver capacities.
## Summary of changes
Add `pageserver_timeline_states_count` with two labels: active +
offloaded.
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
If a shard split fails and must roll back, the tenant may hit a cold
start as the parent shard's files have already been removed from local
disk.
External contribution with minor adjustments, see
https://neondb.slack.com/archives/C08TE3203RQ/p1748246398269309.
## Summary of changes
Keep the parent shard's files on local disk until the split has been
committed, such that they are available if the spilt is rolled back. If
all else fails, the files will be removed on the next Pageserver
restart.
This should also be fine in a mixed version:
* New storcon, old Pageserver: the Pageserver will delete the files
during the split, storcon will log an error when the cleanup detach
fails.
* Old storcon, new Pageserver: the Pageserver will leave the parent's
files around until the next Pageserver restart.
The change looks good to me, but shard splits are delicate so I'd like
some extra eyes on this.
