book/rocksdb
1
0
Fork 0
mirror of https://github.com/facebook/rocksdb.git synced 2025-04-20 18:09:44 +08:00
Code Issues Packages Projects Releases Wiki Activity
rocksdb/unreleased_history/performance_improvements/cra_regression.md
Hui Xiao f69a5fe8ee Cap compaction_readahead_size by max_sectors_kb (#12937) 
Summary:
**Context/Summary:**
https://github.com/facebook/rocksdb/issues/12038 reported a regression where compaction read ahead does not work when `compaction_readahead_size ` is greater than `max_sectors_kb` defined in linux (i.e, largest I/O size that the OS issues to a block device, see https://www.kernel.org/doc/Documentation/block/queue-sysfs.txt for more).

This PR fixes it by capping the `compaction_readahead_size` by `max_sectors_kb` if any. A refactoring of reading queue sys file is also included to reuse existing code.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/12937

Test Plan: https://github.com/facebook/rocksdb/issues/12038#issuecomment-2327618031 verified the regression was fixed

Reviewed By: anand1976

Differential Revision: D61350188

Pulled By: hx235

fbshipit-source-id: e10677f2f5854c22ebf6318b052557db94b98abe
2024-11-18 15:08:21 -08:00

183 B
Raw Blame History

Fix regression in issue #12038 due to Options::compaction_readahead_size greater than max_sectors_kb (i.e, largest I/O size that the OS issues to a block device defined in linux)