-
Notifications
You must be signed in to change notification settings - Fork 3.3k
ReleaseNotes1641
Date: 2026-3-6
http://www.memcached.org/files/memcached-1.6.41.tar.gz
Important bugfix release over regressions starting from 1.6.34 and 1.6.40. The glitches are not likely critical but we want to raise awareness of them.
In 1.6.34 we introduced new memory mover code. In very rare cases a bug can lead to the memory mover becoming unable to make progress and pages will no longer be moved for either a period of time or forever.
This also introduced a rare crash when slabs_mover=2 start option is in
use. However it is highly recommended that nobody use this option as it will
give poor hit rates for most workloads.
In 1.6.40 we introduced a rewritten protocol parser. The lru tuning command
was made inaccessible and is now fixed in this release.
Finally, this repairs a bug in extstore where frequently accessed items can be lost during data compaction. This has existed since extstore was written and is not a recent regression. It is more likely to happen on highly loaded systems with low memory available to keep active items in RAM vs disk.
Several ASAN triggering but otherwise harmless bugs were also fixed.
- tests: make slabs-reassign2 test more resilient
- proxy: reduce flakiness in t/proxyunits.t
- proxy: fix off by one in temp string with 250b key
- slabs: fix hang and crash.
- Fix failing proxy*.t tests on some systems like OL8
- Account for absent 'ssl_proto_errors' in stats during SSL tests
- Fix test compatibility on IPv6-only systems.
- Use SSLv23 method when TLSv1.3 is unsupported (e.g., macOS)
- extstore: more compaction write patience
- parser: fix
lrucommand regression - Fix: avoid null print for slab busy reason
- extstore: testing around rescued compaction items
- extstore: fix compaction checks wrong refcount
- proto: armor against empty commands
None
The following people contributed to this release since 1.6.40.
Note that this is based on who contributed changes, not how they were done. In many cases, a code snippet on the mailing list or a bug report ended up as a commit with your name on it.
Note that this is just a summary of how many changes each person made
which doesn't necessarily reflect how significant each change was.
For details on what led up into a branch, either grab the git repo and
look at the output of git log 1.6.40..1.6.41 or use a web view.
- Repo list: https://github.com/memcached/memcached/wiki/DevelopmentRepos
- Web View: http://github.com/memcached/memcached/commits/1.6.41
10 dormando
6 Bujna, Igor
1 Sunjeet