Avoids compiling the crate and its dependencies into binaries that don't
need them. Shrinks the compute_ctl binary from about 31MB to 28MB in the
release-line-debug-size-lto profile.
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>
# 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.
## 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>
## 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.
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>
## 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
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
Follow-up of #9234 to give hyper 1.0 the version-free name, and the
legacy version of hyper the one with the version number inside. As we
move away from hyper 0.14, we can remove the `hyper0` name piece by
piece.
Part of #9255
This PR simplifies the pageserver configuration parsing as follows:
* introduce the `pageserver_api::config::ConfigToml` type
* implement `Default` for `ConfigToml`
* use serde derive to do the brain-dead leg-work of processing the toml
document
* use `serde(default)` to fill in default values
* in `pageserver` crate:
* use `toml_edit` to deserialize the pageserver.toml string into a
`ConfigToml`
* `PageServerConfig::parse_and_validate` then
* consumes the `ConfigToml`
* destructures it exhaustively into its constituent fields
* constructs the `PageServerConfig`
The rules are:
* in `ConfigToml`, use `deny_unknown_fields` everywhere
* static default values go in `pageserver_api`
* if there cannot be a static default value (e.g. which default IO
engine to use, because it depends on the runtime), make the field in
`ConfigToml` an `Option`
* if runtime-augmentation of a value is needed, do that in
`parse_and_validate`
* a good example is `virtual_file_io_engine` or `l0_flush`, both of
which need to execute code to determine the effective value in
`PageServerConf`
The benefits:
* massive amount of brain-dead repetitive code can be deleted
* "unused variable" compile-time errors when removing a config value,
due to the exhaustive destructuring in `parse_and_validate`
* compile-time errors guide you when adding a new config field
Drawbacks:
* serde derive is sometimes a bit too magical
* `deny_unknown_fields` is easy to miss
Future Work / Benefits:
* make `neon_local` use `pageserver_api` to construct `ConfigToml` and
write it to `pageserver.toml`
* This provides more type safety / coompile-time errors than the current
approach.
### Refs
Fixes#3682
### Future Work
* `remote_storage` deser doesn't reject unknown fields
https://github.com/neondatabase/neon/issues/8915
* clean up `libs/pageserver_api/src/config.rs` further
* break up into multiple files, at least for tenant config
* move `models` as appropriate / refine distinction between config and
API models / be explicit about when it's the same
* use `pub(crate)` visibility on `mod defaults` to detect stale values
In proxy I switched to a leaky-bucket impl using the GCRA algorithm. I
figured I could share the code with pageserver and remove the
leaky_bucket crate dependency with some very basic tokio timers and
queues for fairness.
The underlying algorithm should be fairly clear how it works from the
comments I have left in the code.
---
In benchmarking pageserver, @problame found that the new implementation
fixes a getpage throughput discontinuity in pageserver under the
`pagebench get-page-latest-lsn` benchmark with the clickbench dataset
(`test_perf_olap.py`).
The discontinuity is that for any of `--num-clients={2,3,4}`, getpage
throughput remains 10k.
With `--num-clients=5` and greater, getpage throughput then jumps to the
configured 20k rate limit.
With the changes in this PR, the discontinuity is gone, and we scale
throughput linearly to `--num-clients` until the configured rate limit.
More context in
https://github.com/neondatabase/cloud/issues/16886#issuecomment-2315257641.
closes https://github.com/neondatabase/cloud/issues/16886
---------
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
Part of [Epic: Bypass PageCache for user data
blocks](https://github.com/neondatabase/neon/issues/7386).
# Problem
`InMemoryLayer` still uses the `PageCache` for all data stored in the
`VirtualFile` that underlies the `EphemeralFile`.
# Background
Before this PR, `EphemeralFile` is a fancy and (code-bloated) buffered
writer around a `VirtualFile` that supports `blob_io`.
The `InMemoryLayerInner::index` stores offsets into the `EphemeralFile`.
At those offset, we find a varint length followed by the serialized
`Value`.
Vectored reads (`get_values_reconstruct_data`) are not in fact vectored
- each `Value` that needs to be read is read sequentially.
The `will_init` bit of information which we use to early-exit the
`get_values_reconstruct_data` for a given key is stored in the
serialized `Value`, meaning we have to read & deserialize the `Value`
from the `EphemeralFile`.
The L0 flushing **also** needs to re-determine the `will_init` bit of
information, by deserializing each value during L0 flush.
# Changes
1. Store the value length and `will_init` information in the
`InMemoryLayer::index`. The `EphemeralFile` thus only needs to store the
values.
2. For `get_values_reconstruct_data`:
- Use the in-memory `index` figures out which values need to be read.
Having the `will_init` stored in the index enables us to do that.
- View the EphemeralFile as a byte array of "DIO chunks", each 512 bytes
in size (adjustable constant). A "DIO chunk" is the minimal unit that we
can read under direct IO.
- Figure out which chunks need to be read to retrieve the serialized
bytes for thes values we need to read.
- Coalesce chunk reads such that each DIO chunk is only read once to
serve all value reads that need data from that chunk.
- Merge adjacent chunk reads into larger
`EphemeralFile::read_exact_at_eof_ok` of up to 128k (adjustable
constant).
3. The new `EphemeralFile::read_exact_at_eof_ok` fills the IO buffer
from the underlying VirtualFile and/or its in-memory buffer.
4. The L0 flush code is changed to use the `index` directly, `blob_io`
5. We can remove the `ephemeral_file::page_caching` construct now.
The `get_values_reconstruct_data` changes seem like a bit overkill but
they are necessary so we issue the equivalent amount of read system
calls compared to before this PR where it was highly likely that even if
the first PageCache access was a miss, remaining reads within the same
`get_values_reconstruct_data` call from the same `EphemeralFile` page
were a hit.
The "DIO chunk" stuff is truly unnecessary for page cache bypass, but,
since we're working on [direct
IO](https://github.com/neondatabase/neon/issues/8130) and
https://github.com/neondatabase/neon/issues/8719 specifically, we need
to do _something_ like this anyways in the near future.
# Alternative Design
The original plan was to use the `vectored_blob_io` code it relies on
the invariant of Delta&Image layers that `index order == values order`.
Further, `vectored_blob_io` code's strategy for merging IOs is limited
to adjacent reads. However, with direct IO, there is another level of
merging that should be done, specifically, if multiple reads map to the
same "DIO chunk" (=alignment-requirement-sized and -aligned region of
the file), then it's "free" to read the chunk into an IO buffer and
serve the two reads from that buffer.
=> https://github.com/neondatabase/neon/issues/8719
# Testing / Performance
Correctness of the IO merging code is ensured by unit tests.
Additionally, minimal tests are added for the `EphemeralFile`
implementation and the bit-packed `InMemoryLayerIndexValue`.
Performance testing results are presented below.
All pref testing done on my M2 MacBook Pro, running a Linux VM.
It's a release build without `--features testing`.
We see definitive improvement in ingest performance microbenchmark and
an ad-hoc microbenchmark for getpage against InMemoryLayer.
```
baseline: commit 7c74112b2a origin/main
HEAD: ef1c55c52e
```
<details>
```
cargo bench --bench bench_ingest -- 'ingest 128MB/100b seq, no delta'
baseline
ingest-small-values/ingest 128MB/100b seq, no delta
time: [483.50 ms 498.73 ms 522.53 ms]
thrpt: [244.96 MiB/s 256.65 MiB/s 264.73 MiB/s]
HEAD
ingest-small-values/ingest 128MB/100b seq, no delta
time: [479.22 ms 482.92 ms 487.35 ms]
thrpt: [262.64 MiB/s 265.06 MiB/s 267.10 MiB/s]
```
</details>
We don't have a micro-benchmark for InMemoryLayer and it's quite
cumbersome to add one. So, I did manual testing in `neon_local`.
<details>
```
./target/release/neon_local stop
rm -rf .neon
./target/release/neon_local init
./target/release/neon_local start
./target/release/neon_local tenant create --set-default
./target/release/neon_local endpoint create foo
./target/release/neon_local endpoint start foo
psql 'postgresql://cloud_admin@127.0.0.1:55432/postgres'
psql (13.16 (Debian 13.16-0+deb11u1), server 15.7)
CREATE TABLE wal_test (
id SERIAL PRIMARY KEY,
data TEXT
);
DO $$
DECLARE
i INTEGER := 1;
BEGIN
WHILE i <= 500000 LOOP
INSERT INTO wal_test (data) VALUES ('data');
i := i + 1;
END LOOP;
END $$;
-- => result is one L0 from initdb and one 137M-sized ephemeral-2
DO $$
DECLARE
i INTEGER := 1;
random_id INTEGER;
random_record wal_test%ROWTYPE;
start_time TIMESTAMP := clock_timestamp();
selects_completed INTEGER := 0;
min_id INTEGER := 1; -- Minimum ID value
max_id INTEGER := 100000; -- Maximum ID value, based on your insert range
iters INTEGER := 100000000; -- Number of iterations to run
BEGIN
WHILE i <= iters LOOP
-- Generate a random ID within the known range
random_id := min_id + floor(random() * (max_id - min_id + 1))::int;
-- Select the row with the generated random ID
SELECT * INTO random_record
FROM wal_test
WHERE id = random_id;
-- Increment the select counter
selects_completed := selects_completed + 1;
-- Check if a second has passed
IF EXTRACT(EPOCH FROM clock_timestamp() - start_time) >= 1 THEN
-- Print the number of selects completed in the last second
RAISE NOTICE 'Selects completed in last second: %', selects_completed;
-- Reset counters for the next second
selects_completed := 0;
start_time := clock_timestamp();
END IF;
-- Increment the loop counter
i := i + 1;
END LOOP;
END $$;
./target/release/neon_local stop
baseline: commit 7c74112b2a origin/main
NOTICE: Selects completed in last second: 1864
NOTICE: Selects completed in last second: 1850
NOTICE: Selects completed in last second: 1851
NOTICE: Selects completed in last second: 1918
NOTICE: Selects completed in last second: 1911
NOTICE: Selects completed in last second: 1879
NOTICE: Selects completed in last second: 1858
NOTICE: Selects completed in last second: 1827
NOTICE: Selects completed in last second: 1933
ours
NOTICE: Selects completed in last second: 1915
NOTICE: Selects completed in last second: 1928
NOTICE: Selects completed in last second: 1913
NOTICE: Selects completed in last second: 1932
NOTICE: Selects completed in last second: 1846
NOTICE: Selects completed in last second: 1955
NOTICE: Selects completed in last second: 1991
NOTICE: Selects completed in last second: 1973
```
NB: the ephemeral file sizes differ by ca 1MiB, ours being 1MiB smaller.
</details>
# Rollout
This PR changes the code in-place and is not gated by a feature flag.
## Problem
We lack a rust bench for the inmemory layer and delta layer write paths:
it is useful to benchmark these components independent of postgres & WAL
decoding.
Related: https://github.com/neondatabase/neon/issues/8452
## Summary of changes
- Refactor DeltaLayerWriter to avoid carrying a Timeline, so that it can
be cleanly tested + benched without a Tenant/Timeline test harness. It
only needed the Timeline for building `Layer`, so this can be done in a
separate step.
- Add `bench_ingest`, which exercises a variety of workload "shapes"
(big values, small values, sequential keys, random keys)
- Include a small uncontroversial optimization: in `freeze`, only
exhaustively walk values to assert ordering relative to end_lsn in debug
mode.
These benches are limited by drive performance on a lot of machines, but
still useful as a local tool for iterating on CPU/memory improvements
around this code path.
Anecdotal measurements on Hetzner AX102 (Ryzen 7950xd):
```
ingest-small-values/ingest 128MB/100b seq
time: [1.1160 s 1.1230 s 1.1289 s]
thrpt: [113.38 MiB/s 113.98 MiB/s 114.70 MiB/s]
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) low mild
Benchmarking ingest-small-values/ingest 128MB/100b rand: Warming up for 3.0000 s
Warning: Unable to complete 10 samples in 10.0s. You may wish to increase target time to 18.9s.
ingest-small-values/ingest 128MB/100b rand
time: [1.9001 s 1.9056 s 1.9110 s]
thrpt: [66.982 MiB/s 67.171 MiB/s 67.365 MiB/s]
Benchmarking ingest-small-values/ingest 128MB/100b rand-1024keys: Warming up for 3.0000 s
Warning: Unable to complete 10 samples in 10.0s. You may wish to increase target time to 11.0s.
ingest-small-values/ingest 128MB/100b rand-1024keys
time: [1.0715 s 1.0828 s 1.0937 s]
thrpt: [117.04 MiB/s 118.21 MiB/s 119.46 MiB/s]
ingest-small-values/ingest 128MB/100b seq, no delta
time: [425.49 ms 429.07 ms 432.04 ms]
thrpt: [296.27 MiB/s 298.32 MiB/s 300.83 MiB/s]
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) low mild
ingest-big-values/ingest 128MB/8k seq
time: [373.03 ms 375.84 ms 379.17 ms]
thrpt: [337.58 MiB/s 340.57 MiB/s 343.13 MiB/s]
Found 1 outliers among 10 measurements (10.00%)
1 (10.00%) high mild
ingest-big-values/ingest 128MB/8k seq, no delta
time: [81.534 ms 82.811 ms 83.364 ms]
thrpt: [1.4994 GiB/s 1.5095 GiB/s 1.5331 GiB/s]
Found 1 outliers among 10 measurements (10.00%)
```
## Problem
We recently added a "visibility" state to layers, but nothing
initializes it.
Part of:
- #8398
## Summary of changes
- Add a dependency on `range-set-blaze`, which is used as a fast
incrementally updated alternative to KeySpace. We could also use this to
replace the internals of KeySpaceRandomAccum if we wanted to. Writing a
type that does this kind of "BtreeMap & merge overlapping entries" thing
isn't super complicated, but no reason to write this ourselves when
there's a third party impl available.
- Add a function to layermap to calculate visibilities for each layer
- Add a function to Timeline to call into layermap and then apply these
visibilities to the Layer objects.
- Invoke the calculation during startup, after image layer creations,
and when removing branches. Branch removal and image layer creation are
the two ways that a layer can go from Visible to Covered.
- Add unit test & benchmark for the visibility calculation
- Expose `pageserver_visible_physical_size` metric, which should always
be <= `pageserver_remote_physical_size`.
- This metric will feed into the /v1/utilization endpoint later: the
visible size indicates how much space we would like to use on this
pageserver for this tenant.
- When `pageserver_visible_physical_size` is greater than
`pageserver_resident_physical_size`, this is a sign that the tenant has
long-idle branches, which result in layers that are visible in
principle, but not used in practice.
This does not keep visibility hints up to date in all cases:
particularly, when creating a child timeline, any previously covered
layers will not get marked Visible until they are accessed.
Updates after image layer creation could be implemented as more of a
special case, but this would require more new code: the existing depth
calculation code doesn't maintain+yield the list of deltas that would be
covered by an image layer.
## Performance
This operation is done rarely (at startup and at timeline deletion), so
needs to be efficient but not ultra-fast.
There is a new `visibility` bench that measures runtime for a synthetic
100k layers case (`sequential`) and a real layer map (`real_map`) with
~26k layers.
The benchmark shows runtimes of single digit milliseconds (on a ryzen
7950). This confirms that the runtime shouldn't be a problem at startup
(as we already incur S3-level latencies there), but that it's slow
enough that we definitely shouldn't call it more often than necessary,
and it may be worthwhile to optimize further later (things like: when
removing a branch, only bother scanning layers below the branchpoint)
```
visibility/sequential time: [4.5087 ms 4.5894 ms 4.6775 ms]
change: [+2.0826% +3.9097% +5.8995%] (p = 0.00 < 0.05)
Performance has regressed.
Found 24 outliers among 100 measurements (24.00%)
2 (2.00%) high mild
22 (22.00%) high severe
min: 0/1696070, max: 93/1C0887F0
visibility/real_map time: [7.0796 ms 7.0832 ms 7.0871 ms]
change: [+0.3900% +0.4505% +0.5164%] (p = 0.00 < 0.05)
Change within noise threshold.
Found 4 outliers among 100 measurements (4.00%)
3 (3.00%) high mild
1 (1.00%) high severe
min: 0/1696070, max: 93/1C0887F0
visibility/real_map_many_branches
time: [4.5285 ms 4.5355 ms 4.5434 ms]
change: [-1.0012% -0.8004% -0.5969%] (p = 0.00 < 0.05)
Change within noise threshold.
```
## Problem
- Resident memory on long running pageserver processes tends to climb:
memory fragmentation is suspected.
- Total resident memory may be a limiting factor for running on smaller
nodes.
## Summary of changes
- As a low-energy experiment, switch the pageserver to use jemalloc (not
a net-new dependency, proxy already use it)
- Decide at end of week whether to revert before next release.
Extracted from https://github.com/neondatabase/neon/pull/7375. We assume
everything >= 0x80 are metadata keys. AUX file keys are part of the
metadata keys, and we use `0x90` as the prefix for AUX file keys.
The AUX file encoding is described in the code comment. We use xxhash128
as the hash algorithm. It seems to be portable according to the
introduction,
> xxHash is an Extremely fast Hash algorithm, processing at RAM speed
limits. Code is highly portable, and produces hashes identical across
all platforms (little / big endian).
...though whether the Rust version follows the same convention is
unknown and might need manual review of the library. Anyways, we can
always change the hash algorithm before rolling it out in
staging/end-user, and I made a quick decision to use xxhash here because
it generates 128b hash + portable. We can save the discussion of which
hash algorithm to use later.
---------
Signed-off-by: Alex Chi Z <chi@neon.tech>
## Problem
Follows: https://github.com/neondatabase/neon/pull/7182
- Sufficient concurrent writes could OOM a pageserver from the size of
indices on all the InMemoryLayer instances.
- Enforcement of checkpoint_period only happened if there were some
writes.
Closes: https://github.com/neondatabase/neon/issues/6916
## Summary of changes
- Add `ephemeral_bytes_per_memory_kb` config property. This controls the
ratio of ephemeral layer capacity to memory capacity. The weird unit is
to enable making the ratio less than 1:1 (set this property to 1024 to
use 1MB of ephemeral layers for every 1MB of RAM, set it smaller to get
a fraction).
- Implement background layer rolling checks in
Timeline::compaction_iteration -- this ensures we apply layer rolling
policy in the absence of writes.
- During background checks, if the total ephemeral layer size has
exceeded the limit, then roll layers whose size is greater than the mean
size of all ephemeral layers.
- Remove the tick() path from walreceiver: it isn't needed any more now
that we do equivalent checks from compaction_iteration.
- Add tests for the above.
---------
Co-authored-by: Arpad Müller <arpad-m@users.noreply.github.com>
refs https://github.com/neondatabase/neon/issues/7136
Problem
-------
Before this PR, we were using
`tokio_epoll_uring::thread_local_system()`,
which panics on tokio_epoll_uring::System::launch() failure
As we've learned in [the
past](https://github.com/neondatabase/neon/issues/6373#issuecomment-1905814391),
some older Linux kernels account io_uring instances as locked memory.
And while we've raised the limit in prod considerably, we did hit it
once on 2024-03-11 16:30 UTC.
That was after we enabled tokio-epoll-uring fleet-wide, but before
we had shipped release-5090 (c6ed86d3d0)
which did away with the last mass-creation of tokio-epoll-uring
instances as per
commit 3da410c8fe
Author: Christian Schwarz <christian@neon.tech>
Date: Tue Mar 5 10:03:54 2024 +0100
tokio-epoll-uring: use it on the layer-creating code paths (#6378)
Nonetheless, it highlighted that panicking in this situation is probably
not ideal, as it can leave the pageserver process in a semi-broken
state.
Further, due to low sampling rate of Prometheus metrics, we don't know
much about the circumstances of this failure instance.
Solution
--------
This PR implements a custom thread_local_system() that is
pageserver-aware
and will do the following on failure:
- dump relevant stats to `tracing!`, hopefully they will be useful to
understand the circumstances better
- if it's the locked memory failure (or any other ENOMEM): abort() the
process
- if it's ENOMEM, retry with exponential back-off, capped at 3s.
- add metric counters so we can create an alert
This makes sense in the production environment where we know that
_usually_, there's ample locked memory allowance available, and we know
the failure rate is rare.
Rebased version of #5234, part of #6768
This consists of three parts:
1. A refactoring and new contract for implementing and testing
compaction.
The logic is now in a separate crate, with no dependency on the
'pageserver' crate. It defines an interface that the real pageserver
must implement, in order to call the compaction algorithm. The interface
models things like delta and image layers, but just the parts that the
compaction algorithm needs to make decisions. That makes it easier unit
test the algorithm and experiment with different implementations.
I did not convert the current code to the new abstraction, however. When
compaction algorithm is set to "Legacy", we just use the old code. It
might be worthwhile to convert the old code to the new abstraction, so
that we can compare the behavior of the new algorithm against the old
one, using the same simulated cases. If we do that, have to be careful
that the converted code really is equivalent to the old.
This inclues only trivial changes to the main pageserver code. All the
new code is behind a tenant config option. So this should be pretty safe
to merge, even if the new implementation is buggy, as long as we don't
enable it.
2. A new compaction algorithm, implemented using the new abstraction.
The new algorithm is tiered compaction. It is inspired by the PoC at PR
#4539, although I did not use that code directly, as I needed the new
implementation to fit the new abstraction. The algorithm here is less
advanced, I did not implement partial image layers, for example. I
wanted to keep it simple on purpose, so that as we add bells and
whistles, we can see the effects using the included simulator.
One difference to #4539 and your typical LSM tree implementations is how
we keep track of the LSM tree levels. This PR doesn't have a permanent
concept of a level, tier or sorted run at all. There are just delta and
image layers. However, when compaction starts, we look at the layers
that exist, and arrange them into levels, depending on their shapes.
That is ephemeral: when the compaction finishes, we forget that
information. This allows the new algorithm to work without any extra
bookkeeping. That makes it easier to transition from the old algorithm
to new, and back again.
There is just a new tenant config option to choose the compaction
algorithm. The default is "Legacy", meaning the current algorithm in
'main'. If you set it to "Tiered", the new algorithm is used.
3. A simulator, which implements the new abstraction.
The simulator can be used to analyze write and storage amplification,
without running a test with the full pageserver. It can also draw an SVG
animation of the simulation, to visualize how layers are created and
deleted.
To run the simulator:
cargo run --bin compaction-simulator run-suite
---------
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
The rust stdlib uses the efficient `posix_spawn` by default.
However, before this PR, pageserver used `pre_exec()` in our
`close_fds()` ext trait.
This PR moves the work that `close_fds()` did to the walredo C code.
I verified manually using `gdb` that we're now forking out the walredo
process using `posix_spawn`.
refs https://github.com/neondatabase/neon/issues/6565
Dependency (commits inline):
https://github.com/neondatabase/neon/pull/5842
## Problem
Secondary mode tenants need a manifest of what to download. Ultimately
this will be some kind of heat-scored set of layers, but as a robust
first step we will simply use the set of resident layers: secondary
tenant locations will aim to match the on-disk content of the attached
location.
## Summary of changes
- Add heatmap types representing the remote structure
- Add hooks to Tenant/Timeline for generating these heatmaps
- Create a new `HeatmapUploader` type that is external to `Tenant`, and
responsible for walking the list of attached tenants and scheduling
heatmap uploads.
Notes to reviewers:
- Putting the logic for uploads (and later, secondary mode downloads)
outside of `Tenant` is an opinionated choice, motivated by:
- Enable future smarter scheduling of operations, e.g. uploading the
stalest tenant first, rather than having all tenants compete for a fair
semaphore on a first-come-first-served basis. Similarly for downloads,
we may wish to schedule the tenants with the hottest un-downloaded
layers first.
- Enable accessing upload-related state without synchronization (it
belongs to HeatmapUploader, rather than being some Mutex<>'d part of
Tenant)
- Avoid further expanding the scope of Tenant/Timeline types, which are
already among the largest in the codebase
- You might reasonably wonder how much of the uploader code could be a
generic job manager thing. Probably some of it: but let's defer pulling
that out until we have at least two users (perhaps secondary downloads
will be the second one) to highlight which bits are really generic.
Compromises:
- Later, instead of using digests of heatmaps to decide whether anything
changed, I would prefer to avoid walking the layers in tenants that
don't have changes: tracking that will be a bit invasive, as it needs
input from both remote_timeline_client and Layer.
Remove handcrafted TenantConf deserialization code. Use
`serde_path_to_error` to include the field which failed parsing. Leaves
the duplicated TenantConf in pageserver and models, does not touch
PageserverConf handcrafted deserialization.
Error change:
- before change: "configure option `checkpoint_distance` cannot be
negative"
- after change: "`checkpoint_distance`: invalid value: integer `-1`,
expected u64"
Fixes: #5300
Cc: #3682
---------
Signed-off-by: Rahul Modpur <rmodpur2@gmail.com>
Co-authored-by: Shany Pozin <shany@neon.tech>
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
Fixes#4689 by replacing all of `std::Path` , `std::PathBuf` with
`camino::Utf8Path`, `camino::Utf8PathBuf` in
- pageserver
- safekeeper
- control_plane
- libs/remote_storage
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
Write collected metrics to disk to recover previously sent metrics on
restart.
Recover the previously collected metrics during startup, send them over
at right time
- send cached synthetic size before actual is calculated
- when `last_record_lsn` rolls back on startup
- stay at last sent `written_size` metric
- send `written_size_delta_bytes` metric as 0
Add test support: stateful verification of events in python tests.
Fixes: #5206
Cc: #5175 (loggings, will be enhanced in follow-up)
## Problem
The `EphemeralFile::write_blob` function accesses the page cache
internally. We want to require `async` for these accesses in #5023.
## Summary of changes
This removes the implementaiton of the `BlobWriter` trait for
`EphemeralFile` and turns the `write_blob` function into an inherent
function. We can then make it async as well as the `push_bytes`
function. We move the `SER_BUFFER` thread-local into the
`InMemoryLayerInner` so that the same buffer can be accessed by
different threads as the async is (potentially) moved between threads.
Part of #4743, preparation for #5023.
Adds in a barrier for the duration of the `Tenant::shutdown`.
`pageserver_shutdown` will join this await, `detach`es and `ignore`s
will not.
Fixes#4429.
---------
Co-authored-by: Christian Schwarz <christian@neon.tech>
Introduce read timeouts to our `page_service` connections. Without read
timeouts, we essentially leak connections.
This is a port of #3995. Split the refactorings to the other PR: #4097.
Fixes#4028.
This patch adds a pageserver-global background loop that evicts layers
in response to a shortage of available bytes in the $repo/tenants
directory's filesystem.
The loop runs periodically at a configurable `period`.
Each loop iteration uses `statvfs` to determine filesystem-level space
usage. It compares the returned usage data against two different types
of thresholds. The iteration tries to evict layers until app-internal
accounting says we should be below the thresholds. We cross-check this
internal accounting with the real world by making another `statvfs` at
the end of the iteration. We're good if that second statvfs shows that
we're _actually_ below the configured thresholds. If we're still above
one or more thresholds, we emit a warning log message, leaving it to the
operator to investigate further.
There are two thresholds:
- `max_usage_pct` is the relative available space, expressed in percent
of the total filesystem space. If the actual usage is higher, the
threshold is exceeded.
- `min_avail_bytes` is the absolute available space in bytes. If the
actual usage is lower, the threshold is exceeded.
The iteration evicts layers in LRU fashion with a reservation of up to
`tenant_min_resident_size` bytes of the most recent layers per tenant.
The layers not part of the per-tenant reservation are evicted
least-recently-used first until we're below all thresholds. The
`tenant_min_resident_size` can be overridden per tenant as
`min_resident_size_override` (bytes).
In addition to the loop, there is also an HTTP endpoint to perform one
loop iteration synchronous to the request. The endpoint takes an
absolute number of bytes that the iteration needs to evict before
pressure is relieved. The tests use this endpoint, which is a great
simplification over setting up loopback-mounts in the tests, which would
be required to test the statvfs part of the implementation. We will rely
on manual testing in staging to test the statvfs parts.
The HTTP endpoint is also handy in emergencies where an operator wants
the pageserver to evict a given amount of space _now. Hence, it's
arguments documented in openapi_spec.yml. The response type isn't
documented though because we don't consider it stable. The endpoint
should _not_ be used by Console but it could be used by on-call.
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
Co-authored-by: Dmitry Rodionov <dmitry@neon.tech>
Co-authored-by: Heikki Linnakangas <heikki@neon.tech>
To untie cyclic dependency between sync and async versions of postgres_backend,
copy QueryError and some logging/error routines to postgres_backend.rs. This is
temporal glue to make commits smaller, sync version will be dropped by the
upcoming commit completely.
This patch adds a per-timeline periodic task that executes an eviction
policy. The eviction policy is configurable per tenant.
Two policies exist:
- NoEviction (the default one)
- LayerAccessThreshold
The LayerAccessThreshold policy examines the last access timestamp per
layer in the layer map and evicts the layer if that last access is
further in the past than a configurable threshold value.
This policy kind is evaluated periodically at a configurable period.
It logs a summary statistic at `info!()` or `warn!()` level, depending
on whether any evictions failed.
This feature has no explicit killswitch since it's off by default.
This patch adds basic access statistics for historic layers
and exposes them in the management API's `LayerMapInfo`.
We record the accesses in the `{Delta,Image}Layer::load()` function
because it's the common path of
* page_service (`Timline::get_reconstruct_data()`)
* Compaction (`PersistentLayer::iter()` and `PersistentLayer::key_iter()`)
The stats survive residence status changes, and record these as well.
When scraping the layer map endpoint to record its evolution over time,
one must account for stat resets because they are in-memory only and
will reset on pageserver restart.
Use the launch timestamp header added by (#3527) to identify pageserver restarts.
This is PR https://github.com/neondatabase/neon/pull/3496
It was nice to have and useful at the time, but unfortunately the method
used to gather the profiling data doesn't play nicely with 'async'. PR
#3228 will turn 'get_page_at_lsn' function async, which will break the
profiling support. Let's remove it, and re-introduce some kind of
profiling later, using some different method, if we feel like we need it
again.
The synchronous 'tar' crate has required us to use block_in_place and
SyncIoBridge to work together with the async I/O in the client
connection. Switch to 'tokio-tar' crate that uses async I/O natively.
As part of this, move the CopyDataWriter implementation to
postgres_backend_async.rs. Even though it's only used in one place
currently, it's in principle generally applicable whenever you want to
use COPY out.
Unfortunately we cannot use the 'tokio-tar' as it is: the Builder
implementation requires the writer to have 'static lifetime. So we
have to use a modified version without that requirement. The 'static
lifetime was required just for the Drop implementation that writes
the end-of-archive sections if the Builder is dropped without calling
`finish`. But we don't actually want that behavior anyway; in fact
we had to jump through some hoops with the AbortableWrite hack to skip
those. With the modified version of 'tokio-tar' without that Drop
implementation, we don't need AbortableWrite either.
Co-authored-by: Kirill Bulatov <kirill@neon.tech>
I looked at "cargo tree" output and noticed that through various
dependencies, we are depending on both native-tls and rustls. We have
tried to standardize on rustls for everything, but dependencies on
native-tls have crept in recently. One such dependency came from
'reqwest' with default features in pageserver, used for
consumption_metrics. Another dependency was from 'sentry'. Both
'reqwest' and 'sentry' use native-tls by default, but can use 'rustls'
if compiled with the right feature flags.
Add new background job to collect billing metrics for each tenant and
send them to the HTTP endpoint.
Metrics are cached, so we don't send non-changed metrics.
Add metric collection config parameters:
metric_collection_endpoint (default None, i.e. disabled)
metric_collection_interval (default 60s)
Add test_metric_collection.py to test metric collection
and sending to the mocked HTTP endpoint.
Use port distributor in metric_collection test
review fixes: only update cache after metrics were send successfully, simplify code
disable metric collection if metric_collection_endpoint is not provided in config
This fixes all kinds of problems related to missing params,
like broken timestamps (due to `integer_datetimes`).
This solution is not ideal, but it will help. Meanwhile,
I'm going to dedicate some time to improving connection machinery.
Note that this **does not** fix problems with passing certain parameters
in a reverse direction, i.e. **from client to compute**. This is a
separate matter and will be dealt with in an upcoming PR.
