neon

neon copied to clipboard

Got an interesting case with one of the production read-only endpoints. Walreceiver errored out and died: ``` 2024-06-18 14:09:16.961 {"app":"NeonVM","endpoint_id":"ep-winter-rice-59233042","pod":"compute-lingering-forest-a2yogi5o-6kzkf","_entry":"PG:2024-06-18 14:09:16.598 GMT ttid=a27b300c2ff46c602a1635ab92d236f3/03baf9167a86378faa6375d8273d0f6d sqlstate=53100 [493] FATAL: could not write to...

kelvich

Epic: Safekeepers simulation follow-ups

2

## Motivation https://github.com/neondatabase/neon/pull/5804 should get merged soon. It has the first revision of simulation testing. This epic will contain possible ideas and follow-ups for it. ## Implementation ideas ```[tasklist] ###...

petuhovskiy

Make WAL truncation atomic

1

See https://github.com/zenithdb/zenith/pull/847#discussion_r750131730

arssher

Prototype code: https://github.com/neondatabase/neon/commits/jcsp/sk-concurrency-primitives/ We can use `utils::sync::gate::Gate` to wait for all background tasks to finish before deleting the timeline. What needs to be done before making this change: make sure...

petuhovskiy

safekeeper: remove .partial suffix on the last WAL file.

2

Reasons: - it makes pg_waldump usage slightly more cumbersome, forcing to rename file. - it makes pull_timeline slightly more cumbersome because at any moment source file can be renamed from...

arssher

Save control_file to disk on compute disconnect

1

There was an issue #3317 about lagging `truncate_lsn`. It was mostly fixed, but still can be reproduced with safekeeper restarts: > It's probably still can be reproduced by restarting safekeepers...

petuhovskiy

Use `Timeline::get_vectored` for doing timestamp lookups in SLRUs

2

## Issue Mapping timestamps to LSN is slow for tenats with large SLRU counts. The slowness comes from repeatedly querying SLRUs without using vectored get. On a tenant with 190...

VladLazar

Safekeepers should gossip `remote_consistent_lsn` via broker

1

I saw this happenning in tests: 1. pageserver updates `remote_consistent_lsn` to match `last_record_lsn` 2. connected safekeeper receives new LSN 3. safekeeper updates `broker_is_active` to `false` 4. it stops pushing updates...

petuhovskiy

RFC: move decision-making of desired VM size to VM monitor

2

See also draft implementation of this in: - https://github.com/neondatabase/neon/pull/8113 and - https://github.com/neondatabase/autoscaling/pull/980

hlinnaka

fix(compute_tools): write_postgres_conf missing newline

1

## Problem postgres_conf missing newline ## Summary of changes replace write! with writeln! ## Checklist before requesting a review - [x] I have performed a self-review of my code. -...

fffzlfk