meilisearch / meilisearch/heed

stale lmdb reads on openbsd

Open
#383 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Rust
Stars
914
Forks
82
PR merge metrics
No merged PRs in 30d

Description

On OpenBSD, nested LMDB transactions can read stale data after pages are written through the file descriptor.

Calling msync(..., MS_INVALIDATE) after the writes fixes it. MDB_WRITEMAP also avoids the issue, but cannot be used with nested write transactions.

Tested on OpenBSD 7.9 with heed 0.22.1 and lmdb-master-sys 0.2.6.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the nested-transaction stale-read case on OpenBSD 7.9 with heed 0.22.1 and lmdb-master-sys 0.2.6, then trace the LMDB write and read path. Compare behavior with msync(..., MS_INVALIDATE) and MDB_WRITEMAP; done means nested transactions observe pages written through the file descriptor without relying on the incompatible workaround.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
databases, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.