Investigate using "jemalloc" to see if it helps with memory fragmentation(and potentially performance).
jemalloc is supposed to work well with application that use thread pools(as opposed to constantly creating/destroying threads), and it supposed to reduce memory fragmentation. Seems like a good candidate for 389.
Testing is complete, and results can be seen at:
http://www.port389.org/docs/389ds/FAQ/jemalloc-testing.html
Conclusion:
No regressions observed
Overall memory footprint was significantly smaller compared to the default system malloc. The larger the entry cache was - the larger the magnitude of separation of memory size (especially in the long duration add/delete tests)
Search tests ran faster with jemalloc (~10% improvement). "modify" tests ran for the same amount of time, or just slightly faster.
Metadata Update from @mreynolds: - Issue assigned to mreynolds - Issue set to the milestone: 1.3.4 backlog
389-ds-base is moving from Pagure to Github. This means that new issues and pull requests will be accepted only in 389-ds-base's github repository.
This issue has been cloned to Github and is available here: - https://github.com/389ds/389-ds-base/issues/1460
If you want to receive further updates on the issue, please navigate to the github issue and click on subscribe button.
subscribe
Thank you for understanding. We apologize for all inconvenience.
Metadata Update from @spichugi: - Issue close_status updated to: wontfix (was: Fixed)
Login to comment on this ticket.