Should we also revert the default ReadAdvice.RANDOM setting for 10.2.1 release?
There are multiple reports (1 <https://github.com/apache/lucene/issues/14463>, 2 <https://github.com/apache/lucene/issues/14281>) of performance regressions after this change. Issue #14408 <https://github.com/apache/lucene/issues/14408> has good context on the root cause: a patch in Linux kernel stopped counting MADV_RANDOM accesses towards recency in the page cache's LRU active/inactive list. The commit was in kernel 6.3 but some have reported problems with 6.1 as well, (maybe some kernels backported the patch?). I don't think this RC includes the revert yet. On Thu, Apr 24, 2025 at 9:06 AM Chris Hegarty <christopher.hega...@elastic.co.invalid> wrote: > Please vote for release candidate 1 for Lucene 10.2.1 > > The artifacts can be downloaded from: > > https://dist.apache.org/repos/dist/dev/lucene/lucene-10.2.1-RC1-rev-02cf96b8f7c87fdf5eab864bfc242d36aa2086e7 > > You can run the smoke tester directly with this command: > > python3 -u dev-tools/scripts/smokeTestRelease.py \ > > https://dist.apache.org/repos/dist/dev/lucene/lucene-10.2.1-RC1-rev-02cf96b8f7c87fdf5eab864bfc242d36aa2086e7 > > The vote will be open for at least 72 hours (excluding weekends) i.e. > until 2025-04-29 17:00 UTC. > > [ ] +1 approve > [ ] +0 no opinion > [ ] -1 disapprove (and reason why) > > Here is my +1 > > Draft release notes: > https://cwiki.apache.org/confluence/display/LUCENE/ReleaseNote10_2_1 > > -Chris. > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > > -- - Vigya