## Motivation & Context
We want to move away from `task_mgr` towards explicit tracking of child
tasks.
This PR is extracted from https://github.com/neondatabase/neon/pull/8339
where I refactor `PageRequestHandler` to not depend on task_mgr anymore.
## Changes
This PR refactors all global tasks but `PageRequestHandler` to use some
combination of `JoinHandle`/`JoinSet` + `CancellationToken`.
The `task_mgr::spawn(.., shutdown_process_on_error)` functionality is
preserved through the new `exit_on_panic_or_error` wrapper.
Some global tasks were not using it before, but as of this PR, they are.
The rationale is that all global tasks are relevant for correct
operation of the overall Neon system in one way or another.
## Future Work
After #8339, we can make `task_mgr::spawn` require a `TenantId` instead
of an `Option<TenantId>` which concludes this step of cleanup work and
will help discourage future usage of task_mgr for global tasks.
## Problem
Timelines cannot be deleted if they have children. In many production
cases, a branch or a timeline has been created off the main branch for
various reasons to the effect of having now a "new main" branch. This
feature will make it possible to detach a timeline from its ancestor by
inheriting all of the data before the branchpoint to the detached
timeline and by also reparenting all of the ancestor's earlier branches
to the detached timeline.
## Summary of changes
- Earlier added copy_lsn_prefix functionality is used
- RemoteTimelineClient learns to adopt layers by copying them from
another timeline
- LayerManager adds support for adding adopted layers
-
`timeline::Timeline::{prepare_to_detach,complete_detaching}_from_ancestor`
and `timeline::detach_ancestor` are added
- HTTP PUT handler
Cc: #6994
Co-authored-by: Christian Schwarz <christian@neon.tech>
## Problem
In testing of the earlier fix for OOMs under heavy write load
(https://github.com/neondatabase/neon/pull/7218), we saw that the limit
on ephemeral layer size wasn't being reliably enforced. That was
diagnosed as being due to overwhelmed compaction loops: most tenants
were waiting on the semaphore for background tasks, and thereby not
running the function that proactively rolls layers frequently enough.
Related: https://github.com/neondatabase/neon/issues/6939
## Summary of changes
- Create a new per-tenant background loop for "ingest housekeeping",
which invokes maybe_freeze_ephemeral_layer() without taking the
background task semaphore.
- Downgrade to DEBUG a log line in maybe_freeze_ephemeral_layer that had
been INFO, but turns out to be pretty common in the field.
There's some discussion on the issue
(https://github.com/neondatabase/neon/issues/6939#issuecomment-2083554275)
about alternatives for calling this maybe_freeze_epemeral_layer
periodically without it getting stuck behind compaction. A whole task
just for this feels like kind of a big hammer, but we may in future find
that there are other pieces of lightweight housekeeping that we want to
do here too.
Why is it okay to call maybe_freeze_ephemeral_layer outside of the
background tasks semaphore?
- this is the same work we would do anyway if we receive writes from the
safekeeper, just done a bit sooner.
- The period of the new task is generously jittered (+/- 5%), so when
the ephemeral layer size tips over the threshold, we shouldn't see an
excessively aggressive thundering herd of layer freezes (and only layers
larger than the mean layer size will be frozen)
- All that said, this is an imperfect approach that relies on having a
generous amount of RAM to dip into when we need to freeze somewhat
urgently. It would be nice in future to also block compaction/GC when we
recognize resource stress and need to do other work (like layer
freezing) to reduce memory footprint.
part of https://github.com/neondatabase/neon/issues/7124
Changes
-------
This PR replaces the `EphemeralFile::write_blob`-specifc `struct Writer`
with re-use of `owned_buffers_io::write::BufferedWriter`.
Further, it restructures the code to cleanly separate
* the high-level aspect of EphemeralFile's write_blob / read_blk API
* the page-caching aspect
* the aspect of IO
* performing buffered write IO to an underlying VirtualFile
* serving reads from either the VirtualFile or the buffer if it hasn't
been flushed yet
* the annoying "feature" that reads past the end of the written range
are allowed and expected to return zeroed memory, as long as one remains
within one PAGE_SZ
This PR is an off-by-default revision v2 of the (since-reverted) PR
#6555 / commit `3220f830b7fbb785d6db8a93775f46314f10a99b`.
See that PR for details on why running with a single runtime is
desirable and why we should be ready.
We reverted #6555 because it showed regressions in prodlike cloudbench,
see the revert commit message `ad072de4209193fd21314cf7f03f14df4fa55eb1`
for more context.
This PR makes it an opt-in choice via an env var.
The default is to use the 4 separate runtimes that we have today, there
shouldn't be any performance change.
I tested manually that the env var & added metric works.
```
# undefined env var => no change to before this PR, uses 4 runtimes
./target/debug/neon_local start
# defining the env var enables one-runtime mode, value defines that one runtime's configuration
NEON_PAGESERVER_USE_ONE_RUNTIME=current_thread ./target/debug/neon_local start
NEON_PAGESERVER_USE_ONE_RUNTIME=multi_thread:1 ./target/debug/neon_local start
NEON_PAGESERVER_USE_ONE_RUNTIME=multi_thread:2 ./target/debug/neon_local start
NEON_PAGESERVER_USE_ONE_RUNTIME=multi_thread:default ./target/debug/neon_local start
```
I want to use this change to do more manualy testing and potentially
testing in staging.
Future Work
-----------
Testing / deployment ergonomics would be better if this were a variable
in `pageserver.toml`.
It can be done, but, I don't need it right now, so let's stick with the
env var.
We want to move the code base away from task_mgr.
This PR refactors the walreceiver code such that it doesn't use
`task_mgr` anymore.
# Background
As a reminder, there are three tasks in a Timeline that's ingesting WAL.
`WalReceiverManager`, `WalReceiverConnectionHandler`, and
`WalReceiverConnectionPoller`.
See the documentation in `task_mgr.rs` for how they interact.
Before this PR, cancellation was requested through
task_mgr::shutdown_token() and `TaskHandle::shutdown`.
Wait-for-task-finish was implemented using a mixture of
`task_mgr::shutdown_tasks` and `TaskHandle::shutdown`.
This drawing might help:
<img width="300" alt="image"
src="https://github.com/neondatabase/neon/assets/956573/b6be7ad6-ecb3-41d0-b410-ec85cb8d6d20">
# Changes
For cancellation, the entire WalReceiver task tree now has a
`child_token()` of `Timeline::cancel`. The `TaskHandle` no longer is a
cancellation root.
This means that `Timeline::cancel.cancel()` is propagated.
For wait-for-task-finish, all three tasks in the task tree hold the
`Timeline::gate` open until they exit.
The downside of using the `Timeline::gate` is that we can no longer wait
for just the walreceiver to shut down, which is particularly relevant
for `Timeline::flush_and_shutdown`.
Effectively, it means that we might ingest more WAL while the
`freeze_and_flush()` call is ongoing.
Also, drive-by-fix the assertiosn around task kinds in `wait_lsn`. The
check for `WalReceiverConnectionHandler` was ineffective because that
never was a task_mgr task, but a TaskHandle task. Refine the assertion
to check whether we would wait, and only fail in that case.
# Alternatives
I contemplated (ab-)using the `Gate` by having a separate `Gate` for
`struct WalReceiver`.
All the child tasks would use _that_ gate instead of `Timeline::gate`.
And `struct WalReceiver` itself would hold an `Option<GateGuard>` of the
`Timeline::gate`.
Then we could have a `WalReceiver::stop` function that closes the
WalReceiver's gate, then drops the `WalReceiver::Option<GateGuard>`.
However, such design would mean sharing the WalReceiver's `Gate` in an
`Arc`, which seems awkward.
A proper abstraction would be to make gates hierarchical, analogous to
CancellationToken.
In the end, @jcsp and I talked it over and we determined that it's not
worth the effort at this time.
# Refs
part of #7062
Before this PR, each core had 3 executor threads from 3 different
runtimes. With this PR, we just have one runtime, with one thread per
core. Switching to a single tokio runtime should reduce that effective
over-commit of CPU and in theory help with tail latencies -- iff all
tokio tasks are well-behaved and yield to the runtime regularly.
Are All Tasks Well-Behaved? Are We Ready?
-----------------------------------------
Sadly there doesn't seem to be good out-of-the box tokio tooling to
answer this question.
We *believe* all tasks are well behaved in today's code base, as of the
switch to `virtual_file_io_engine = "tokio-epoll-uring"` in production
(https://github.com/neondatabase/aws/pull/1121).
The only remaining executor-thread-blocking code is walredo and some
filesystem namespace operations.
Filesystem namespace operations work is being tracked in #6663 and not
considered likely to actually block at this time.
Regarding walredo, it currently does a blocking `poll` for read/write to
the pipe file descriptors we use for IPC with the walredo process.
There is an ongoing experiment to make walredo async (#6628), but it
needs more time because there are surprisingly tricky trade-offs that
are articulated in that PR's description (which itself is still WIP).
What's relevant for *this* PR is that
1. walredo is always CPU-bound
2. production tail latencies for walredo request-response
(`pageserver_wal_redo_seconds_bucket`) are
- p90: with few exceptions, low hundreds of micro-seconds
- p95: except on very packed pageservers, below 1ms
- p99: all below 50ms, vast majority below 1ms
- p99.9: almost all around 50ms, rarely at >= 70ms
- [Dashboard
Link](https://neonprod.grafana.net/d/edgggcrmki3uof/2024-03-walredo-latency?orgId=1&var-ds=ZNX49CDVz&var-pXX_by_instance=0.9&var-pXX_by_instance=0.99&var-pXX_by_instance=0.95&var-adhoc=instance%7C%21%3D%7Cpageserver-30.us-west-2.aws.neon.tech&var-per_instance_pXX_max_seconds=0.0005&from=1711049688777&to=1711136088777)
The ones below 1ms are below our current threshold for when we start
thinking about yielding to the executor.
The tens of milliseconds stalls aren't great, but, not least because of
the implicit overcommit of CPU by the three runtimes, we can't be sure
whether these tens of milliseconds are inherently necessary to do the
walredo work or whether we could be faster if there was less contention
for CPU.
On the first item (walredo being always CPU-bound work): it means that
walredo processes will always compete with the executor threads.
We could yield, using async walredo, but then we hit the trade-offs
explained in that PR.
tl;dr: the risk of stalling executor threads through blocking walredo
seems low, and switching to one runtime cleans up one potential source
for higher-than-necessary stall times (explained in the previous
paragraphs).
Code Changes
------------
- Remove the 3 different runtime definitions.
- Add a new definition called `THE_RUNTIME`.
- Use it in all places that previously used one of the 3 removed
runtimes.
- Remove the argument from `task_mgr`.
- Fix failpoint usage where `pausable_failpoint!` should have been used.
We encountered some actual failures because of this, e.g., hung
`get_metric()` calls during test teardown that would client-timeout
after 300s.
As indicated by the comment above `THE_RUNTIME`, we could take this
clean-up further.
But before we create so much churn, let's first validate that there's no
perf regression.
Performance
-----------
We will test this in staging using the various nightly benchmark runs.
However, the worst-case impact of this change is likely compaction
(=>image layer creation) competing with compute requests.
Image layer creation work can't be easily generated & repeated quickly
by pagebench.
So, we'll simply watch getpage & basebackup tail latencies in staging.
Additionally, I have done manual benchmarking using pagebench.
Report:
https://neondatabase.notion.site/2024-03-23-oneruntime-change-benchmarking-22a399c411e24399a73311115fb703ec?pvs=4
Tail latencies and throughput are marginally better (no regression =
good).
Except in a workload with 128 clients against one tenant.
There, the p99.9 and p99.99 getpage latency is about 2x worse (at
slightly lower throughput).
A dip in throughput every 20s (compaction_period_ is clearly visible,
and probably responsible for that worse tail latency.
This has potential to improve with async walredo, and is an edge case
workload anyway.
Future Work
-----------
1. Once this change has shown satisfying results in production, change
the codebase to use the ambient runtime instead of explicitly
referencing `THE_RUNTIME`.
2. Have a mode where we run with a single-threaded runtime, so we
uncover executor stalls more quickly.
3. Switch or write our own failpoints library that is async-native:
https://github.com/neondatabase/neon/issues/7216
## Problem
Tenant deletion had a couple of TODOs where we weren't using proper
cancellation tokens that would have aborted the deletions during process
shutdown.
## Summary of changes
- Refactor enough that deletion/shutdown code has access to the
TenantManager's cancellation toke
- Use that cancellation token in tenant deletion instead of dummy
tokens.
## Problem
Before this PR, it was possible that on-demand downloads were started
after `Timeline::shutdown()`.
For example, we have observed a walreceiver-connection-handler-initiated
on-demand download that was started after `Timeline::shutdown()`s final
`task_mgr::shutdown_tasks()` call.
The underlying issue is that `task_mgr::shutdown_tasks()` isn't sticky,
i.e., new tasks can be spawned during or after
`task_mgr::shutdown_tasks()`.
Cc: https://github.com/neondatabase/neon/issues/4175 in lieu of a more
specific issue for task_mgr. We already decided we want to get rid of it
anyways.
Original investigation:
https://neondb.slack.com/archives/C033RQ5SPDH/p1709824952465949
## Changes
- enter gate while downloading
- use timeline cancellation token for cancelling download
thereby, fixes#7054
Entering the gate might also remove recent "kept the gate from closing"
in staging.
This PR is preliminary cleanups and refactoring around `remote_storage`
for next PR which will move the timeouts and cancellation into
`remote_storage`.
Summary:
- smaller drive-by fixes
- code simplification
- refactor common parts like `DownloadError::is_permanent`
- align error types with `RemoteStorage::list_*` to use more
`download_retry` helper
Cc: #6096
Compaction was holding back timeline deletion because the compaction
lock had been acquired, but the semaphore was waited on. Timeline
deletion was waiting on the same lock for 1500s.
This replaces the
`pageserver::tenant::tasks::concurrent_background_tasks_rate_limit`
(which looks correct) with a simpler `..._permit` which is just an
infallible acquire, which is easier to spot "aah this needs to be raced
with cancellation tokens".
Ref: https://neondb.slack.com/archives/C03F5SM1N02/p1702496912904719
Ref: https://neondb.slack.com/archives/C03F5SM1N02/p1702578093497779
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.
## Problem
- `shutdown_tasks` would log when a particular task was taking a long
time to shut down, but not when it eventually completed. That left one
uncertain as to whether the slow task was the source of a hang, or just
a precursor.
## Summary of changes
- Add a log line after a slow task shutdown
- Add an equivalent in Gate's `warn_if_stuck`, in case we ever need it.
This isn't related to the original issue but was noticed when checking
through these logging paths.
## Problem
In https://github.com/neondatabase/neon/pull/5957, the most essential
types were updated to use TenantShardId rather than TenantId. That
unblocked other work, but didn't fully enable running multiple shards
from the same tenant on the same pageserver.
## Summary of changes
- Use TenantShardId in page cache key for materialized pages
- Update mgr.rs get_tenant() and list_tenants() functions to use a shard
id, and update all callers.
- Eliminate the exactly_one_or_none helper in mgr.rs and all code that
used it
- Convert timeline HTTP routes to use tenant_shard_id
Note on page cache:
```
struct MaterializedPageHashKey {
/// Why is this TenantShardId rather than TenantId?
///
/// Usually, the materialized value of a page@lsn is identical on any shard in the same tenant. However, this
/// this not the case for certain internally-generated pages (e.g. relation sizes). In future, we may make this
/// key smaller by omitting the shard, if we ensure that reads to such pages always skip the cache, or are
/// special-cased in some other way.
tenant_shard_id: TenantShardId,
timeline_id: TimelineId,
key: Key,
}
```
## Problem
When shutting down a Tenant, it isn't just important to cause any
background tasks to stop. It's also important to wait until they have
stopped before declaring shutdown complete, in cases where we may re-use
the tenant's local storage for something else, such as running in
secondary mode, or creating a new tenant with the same ID.
## Summary of changes
A `Gate` class is added, inspired by
[seastar::gate](https://docs.seastar.io/master/classseastar_1_1gate.html).
For types that have an important lifetime that corresponds to some
physical resource, use of a Gate as well as a CancellationToken provides
a robust pattern for async requests & shutdown:
- Requests must always acquire the gate as long as they are using the
object
- Shutdown must set the cancellation token, and then `close()` the gate
to wait for requests in progress before returning.
This is not for memory safety: it's for expressing the difference
between "Arc<Tenant> exists", and "This tenant's files on disk are
eligible to be read/written".
- Both Tenant and Timeline get a Gate & CancellationToken.
- The Timeline gate is held during eviction of layers, and during
page_service requests.
- Existing cancellation support in page_service is refined to use the
timeline-scope cancellation token instead of a process-scope
cancellation token. This replaces the use of `task_mgr::associate_with`:
tasks no longer change their tenant/timelineidentity after being
spawned.
The Tenant's Gate is not yet used, but will be important for
Tenant-scoped operations in secondary mode, where we must ensure that
our secondary-mode downloads for a tenant are gated wrt the activity of
an attached Tenant.
This is part of a broader move away from using the global-state driven
`task_mgr` shutdown tokens:
- less global state where we rely on implicit knowledge of what task a
given function is running in, and more explicit references to the
cancellation token that a particular function/type will respect, making
shutdown easier to reason about.
- eventually avoid the big global TASKS mutex.
---------
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
## Problem
Pageservers must not delete objects or advertise updates to
remote_consistent_lsn without checking that they hold the latest
generation for the tenant in question (see [the RFC](
https://github.com/neondatabase/neon/blob/main/docs/rfcs/025-generation-numbers.md))
In this PR:
- A new "deletion queue" subsystem is introduced, through which
deletions flow
- `RemoteTimelineClient` is modified to send deletions through the
deletion queue:
- For GC & compaction, deletions flow through the full generation
verifying process
- For timeline deletions, deletions take a fast path that bypasses
generation verification
- The `last_uploaded_consistent_lsn` value in `UploadQueue` is replaced
with a mechanism that maintains a "projected" lsn (equivalent to the
previous property), and a "visible" LSN (which is the one that we may
share with safekeepers).
- Until `control_plane_api` is set, all deletions skip generation
validation
- Tests are introduced for the new functionality in
`test_pageserver_generations.py`
Once this lands, if a pageserver is configured with the
`control_plane_api` configuration added in
https://github.com/neondatabase/neon/pull/5163, it becomes safe to
attach a tenant to multiple pageservers concurrently.
---------
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
This PR adds a `task_kind` label to page cache access metrics.
These are to validate our hypothesis that the high hit page cache rate
we observe in prod is due to internal tasks, not getpage requests from
compute.
We believe the latter should near-always be a pageserver-page-cache
_miss_ because compute has it's own page cache, and hence there is no
locality of reference for its accesses to pageserver page cache.
Before this PR, we didn't have `RequestContext` propagation to any code
below the on-demand downloader.
The vast majority of changes in this PR is concerned with adding that
propagation.
Removes a bunch of cases which used `tokio::select` to emulate the
`tokio::time::timeout` function. I've done an additional review on the
cancellation safety of these futures, all of them seem to be
cancellation safe (not that `select!` allows non-cancellation-safe
futures, but as we touch them, such a review makes sense).
Furthermore, I correct a few mentions of a non-existent
`tokio::timeout!` macro in the docs to the `tokio::time::timeout`
function.
## Problem
`cargo +nightly doc` is giving a lot of warnings: broken links, naked
URLs, etc.
## Summary of changes
* update the `proc-macro2` dependency so that it can compile on latest
Rust nightly, see https://github.com/dtolnay/proc-macro2/pull/391 and
https://github.com/dtolnay/proc-macro2/issues/398
* allow the `private_intra_doc_links` lint, as linking to something
that's private is always more useful than just mentioning it without a
link: if the link breaks in the future, at least there is a warning due
to that. Also, one might enable
[`--document-private-items`](https://doc.rust-lang.org/cargo/commands/cargo-doc.html#documentation-options)
in the future and make these links work in general.
* fix all the remaining warnings given by `cargo +nightly doc`
* make it possible to run `cargo doc` on stable Rust by updating
`opentelemetry` and associated crates to version 0.19, pulling in a fix
that previously broke `cargo doc` on stable:
https://github.com/open-telemetry/opentelemetry-rust/pull/904
* Add `cargo doc` to CI to ensure that it won't get broken in the
future.
Fixes#2557
## Future work
* Potentially, it might make sense, for development purposes, to publish
the generated rustdocs somewhere, like for example [how the rust
compiler does
it](https://doc.rust-lang.org/nightly/nightly-rustc/rustc_driver/index.html).
I will file an issue for discussion.
Delete data from s3 when timeline deletion is requested
## Summary of changes
UploadQueue is altered to support scheduling of delete operations in
stopped state. This looks weird, and I'm thinking whether there are
better options/refactorings for upload client to make it look better.
Probably can be part of https://github.com/neondatabase/neon/issues/4378
Deletion is implemented directly in existing endpoint because changes are not
that significant. If we want more safety we can separate those or create
feature flag for new behavior.
resolves [#4193](https://github.com/neondatabase/neon/issues/4193)
---------
Co-authored-by: Joonas Koivunen <joonas@neon.tech>
We have 2 ways of tenant shutdown, we should have just one.
Changes are mostly mechanical simple refactorings.
Added `warn!` on the "shutdown all remaining tasks" should trigger test
failures in the between time of not having solved the "tenant/timeline
owns all spawned tasks" issue.
Cc: #4327.
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>
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
Motivation
==========
Layer Eviction Needs Context
----------------------------
Before we start implementing layer eviction, we need to collect some
access statistics per layer file or maybe even page.
Part of these statistics should be the initiator of a page read request
to answer the question of whether it was page_service vs. one of the
background loops, and if the latter, which of them?
Further, it would be nice to learn more about what activity in the pageserver
initiated an on-demand download of a layer file.
We will use this information to test out layer eviction policies.
Read more about the current plan for layer eviction here:
https://github.com/neondatabase/neon/issues/2476#issuecomment-1370822104
task_mgr problems + cancellation + tenant/timeline lifecycle
------------------------------------------------------------
Apart from layer eviction, we have long-standing problems with task_mgr,
task cancellation, and various races around tenant / timeline lifecycle
transitions.
One approach to solve these is to abandon task_mgr in favor of a
mechanism similar to Golang's context.Context, albeit extended to
support waiting for completion, and specialized to the needs in the
pageserver.
Heikki solves all of the above at once in PR
https://github.com/neondatabase/neon/pull/3228 , which is not yet
merged at the time of writing.
What Is This Patch About
========================
This patch addresses the immediate needs of layer eviction by
introducing a `RequestContext` structure that is plumbed through the
pageserver - all the way from the various entrypoints (page_service,
management API, tenant background loops) down to
Timeline::{get,get_reconstruct_data}.
The struct carries a description of the kind of activity that initiated
the call. We re-use task_mgr::TaskKind for this.
Also, it carries the desired on-demand download behavior of the entrypoint.
Timeline::get_reconstruct_data can then log the TaskKind that initiated
the on-demand download.
I developed this patch by git-checking-out Heikki's big RequestContext
PR https://github.com/neondatabase/neon/pull/3228 , then deleting all
the functionality that we do not need to address the needs for layer
eviction.
After that, I added a few things on top:
1. The concept of attached_child and detached_child in preparation for
cancellation signalling through RequestContext, which will be added in
a future patch.
2. A kill switch to turn DownloadBehavior::Error into a warning.
3. Renamed WalReceiverConnection to WalReceiverConnectionPoller and
added an additional TaskKind WalReceiverConnectionHandler.These were
necessary to create proper detached_child-type RequestContexts for the
various tasks that walreceiver starts.
How To Review This Patch
========================
Start your review with the module-level comment in context.rs.
It explains the idea of RequestContext, what parts of it are implemented
in this patch, and the future plans for RequestContext.
Then review the various `task_mgr::spawn` call sites. At each of them,
we should be creating a new detached_child RequestContext.
Then review the (few) RequestContext::attached_child call sites and
ensure that the spawned tasks do not outlive the task that spawns them.
If they do, these call sites should use detached_child() instead.
Then review the todo_child() call sites and judge whether it's worth the
trouble of plumbing through a parent context from the caller(s).
Lastly, go through the bulk of mechanical changes that simply forwards
the &ctx.
The code in this change was extracted from #2595 (Heikki’s on-demand
download draft PR).
High-Level Changes
- New RemoteLayer Type
- On-Demand Download As An Effect Of Page Reconstruction
- Breaking Semantics For Physical Size Metrics
There are several follow-up work items planned.
Refer to the Epic issue on GitHub: https://github.com/neondatabase/neon/issues/2029
closes https://github.com/neondatabase/neon/pull/3013
Co-authored-by: Kirill Bulatov <kirill@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
New RemoteLayer Type
====================
Instead of downloading all layers during tenant attach, we create
RemoteLayer instances for each of them and add them to the layer map.
On-Demand Download As An Effect Of Page Reconstruction
======================================================
At the heart of pageserver is Timeline::get_reconstruct_data(). It
traverses the layer map until it has collected all the data it needs to
produce the page image. Most code in the code base uses it, though many
layers of indirection.
Before this patch, the function would use synchronous filesystem IO to
load data from disk-resident layer files if the data was not cached.
That is not possible with RemoteLayer, because the layer file has not
been downloaded yet. So, we do the download when get_reconstruct_data
gets there, i.e., “on demand”.
The mechanics of how the download is done are rather involved, because
of the infamous async-sync-async sandwich problem that plagues the async
Rust world. We use the new PageReconstructResult type to work around
this. Its introduction is the cause for a good amount of code churn in
this patch. Refer to the block comment on `with_ondemand_download()`
for details.
Breaking Semantics For Physical Size Metrics
============================================
We rename prometheus metric pageserver_{current,resident}_physical_size to
reflect what this metric actually represents with on-demand download.
This intentionally BREAKS existing grafana dashboard and the cost model data
pipeline. Breaking is desirable because the meaning of this metrics has changed
with on-demand download. See
https://docs.google.com/document/d/12AFpvKY-7FZdR5a4CaD6Ir_rI3QokdCLSPJ6upHxJBo/edit#
for how we will handle this breakage.
Likewise, we rename the new billing_metrics’s PhysicalSize => ResidentSize.
This is not yet used anywhere, so, this is not a breaking change.
There is still a field called TimelineInfo::current_physical_size. It
is now the sum of the layer sizes in layer map, regardless of whether
local or remote. To compute that sum, we added a new trait method
PersistentLayer::file_size().
When updating the Python tests, we got rid of
current_physical_size_non_incremental. An earlier commit removed it from
the OpenAPI spec already, so this is not a breaking change.
test_timeline_size.py has grown additional assertions on the
resident_physical_size metric.
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 should help us in the future to have more freedom with spawning
tasks and cancelling things, most importantly blocking tasks (assuming
the CancellationToken::is_cancelled is performant enough).
CancellationToken allows creation of hierarchical cancellations, which
would also simplify the task_mgr shutdown operation, rendering it
unnecessary.
The code in this change was extracted from PR #2595, i.e., Heikki’s draft
PR for on-demand download.
High-Level Changes
- storage_sync module rewrite
- Changes to Tenant Loading
- Changes to Timeline States
- Crash-safe & Resumable Tenant Attach
There are several follow-up work items planned.
Refer to the Epic issue on GitHub:
https://github.com/neondatabase/neon/issues/2029
Metadata:
closes https://github.com/neondatabase/neon/pull/2785
unsquashed history of this patch: archive/pr-2785-storage-sync2/pre-squash
Co-authored-by: Dmitry Rodionov <dmitry@neon.tech>
Co-authored-by: Christian Schwarz <christian@neon.tech>
===============================================================================
storage_sync module rewrite
===========================
The storage_sync code is rewritten. New module name is storage_sync2, mostly to
make a more reasonable git diff.
The updated block comment in storage_sync2.rs describes the changes quite well,
so, we will not reproduce that comment here. TL;DR:
- Global sync queue and RemoteIndex are replaced with per-timeline
`RemoteTimelineClient` structure that contains a queue for UploadOperations
to ensure proper ordering and necessary metadata.
- Before deleting local layer files, wait for ongoing UploadOps to finish
(wait_completion()).
- Download operations are not queued and executed immediately.
Changes to Tenant Loading
=========================
Initial sync part was rewritten as well and represents the other major change
that serves as a foundation for on-demand downloads. Routines for attaching and
loading shifted directly to Tenant struct and now are asynchronous and spawned
into the background.
Since this patch doesn’t introduce on-demand download of layers we fully
synchronize with the remote during pageserver startup. See details in
`Timeline::reconcile_with_remote` and `Timeline::download_missing`.
Changes to Tenant States
========================
The “Active” state has lost its “background_jobs_running: bool” member. That
variable indicated whether the GC & Compaction background loops are spawned or
not. With this patch, they are now always spawned. Unit tests (#[test]) use the
TenantConf::{gc_period,compaction_period} to disable their effect (15db566).
This patch introduces a new tenant state, “Attaching”. A tenant that is being
attached starts in this state and transitions to “Active” once it finishes
download.
The `GET /tenant` endpoints returns `TenantInfo::has_in_progress_downloads`. We
derive the value for that field from the tenant state now, to remain
backwards-compatible with cloud.git. We will remove that field when we switch
to on-demand downloads.
Changes to Timeline States
==========================
The TimelineInfo::awaits_download field is now equivalent to the tenant being
in Attaching state. Previously, download progress was tracked per timeline.
With this change, it’s only tracked per tenant. When on-demand downloads
arrive, the field will be completely obsolete. Deprecation is tracked in
isuse #2930.
Crash-safe & Resumable Tenant Attach
====================================
Previously, the attach operation was not persistent. I.e., when tenant attach
was interrupted by a crash, the pageserver would not continue attaching after
pageserver restart. In fact, the half-finished tenant directory on disk would
simply be skipped by tenant_mgr because it lacked the metadata file (it’s
written last). This patch introduces an “attaching” marker file inside that is
present inside the tenant directory while the tenant is attaching. During
pageserver startup, tenant_mgr will resume attach if that file is present. If
not, it assumes that the local tenant state is consistent and tries to load the
tenant. If that fails, the tenant transitions into Broken state.
Instead of spawning helper threads, we now use Tokio tasks. There
are multiple Tokio runtimes, for different kinds of tasks. One for
serving libpq client connections, another for background operations
like GC and compaction, and so on. That's not strictly required, we
could use just one runtime, but with this you can still get an
overview of what's happening with "top -H".
There's one subtle behavior in how TenantState is updated. Before this
patch, if you deleted all timelines from a tenant, its GC and
compaction loops were stopped, and the tenant went back to Idle
state. We no longer do that. The empty tenant stays Active. The
changes to test_tenant_tasks.py are related to that.
There's still plenty of synchronous code and blocking. For example, we
still use blocking std::io functions for all file I/O, and the
communication with WAL redo processes is still uses low-level unix
poll(). We might want to rewrite those later, but this will do for
now. The model is that local file I/O is considered to be fast enough
that blocking - and preventing other tasks running in the same thread -
is acceptable.
