neon

neon copied to clipboard

Use smgrexists() instead of access() to enforce uniqueness of generated relfilenumber

3

## Problem Postgres is using `access()` function in `GetNewRelFileNumber` to check if assigned relfilenumber is not used for any other relation. This check will not work in Neon, because we...

knizhnik

feat(pageserver): add k-merge image layer iterator with lazy loading

1

## Problem https://github.com/neondatabase/neon/issues/8002 ## Summary of changes ## Checklist before requesting a review - [ ] I have performed a self-review of my code. - [ ] If it is...

skyzh

https://github.com/neondatabase/neon/pull/8103 adds bottom-most compaction for this unit test. We should use an LSN for the GC horizon instead of time to make the result comparable across runs.

skyzh

RelFileNumber reuse protection is broken

3

GetNewRelFileNumber() checks the filesystem directly with `access()`, to see if a relfilenumber is already in use. That doesn't work with Neon storage, because the relations don't have files in local...

hlinnaka

Epic: Compute Image with all supported versions of Postgres

3

## Motivation It's hard to manage compute pools for 3 Postgres versions. (we have a compute image for each version of Postgres (currently, it's 3 for neonVM and 3 for...

bayandin

LFC: using rolling hyperloglog for correct working set estimation

10

The current LFC uses hyperloglog to estimate the working set size, which does not correctly deal with deletes, and does not capture the working set size used in a window...

skyzh

This Pull Request is created automatically to run the CI pipeline for #8149 Please do not alter or merge/close it. Feel free to review/comment/discuss the original PR #8149.

vipvap

Fix nightly warnings 2024 june

1

## Problem new warnings on nightly. ## Summary of changes broken up each commit by warning type. 1. Remove some unnecessary refs. 2. In edition 2024, inference will default to...

conradludgate

Rough roll-out plan: 1. Switch `--enable-offload` in staging regions, observe for ~week 2. Switch `--enable-offload` in prod regions one by one, observe for ~week 3. Switch `--delete-offloaded-wal` in staging regions,...

petuhovskiy

We should move away from calling our current WAL uploading to remote storage like a "backup" in code/comments. The segment in remote storage can be the only copy of the...

petuhovskiy