Details
-
Bug
-
Resolution: Fixed
-
Major
-
7.6.0
-
Triaged
-
0
-
Yes
Description
We are seeing the same behaviour as https://issues.couchbase.com/browse/MB-58087 for memcached failure tests where failure detection time uses the default time (2min) instead of the value set for timeout.
The regression is coming from build 1352 which contains these commits:
ns_server changed from 0408af8432916f2a160aca8722d9c961a6ac13bf to master
|
[+] 5ebfe79d9 MB-55578: Store MonitorState correctly in health_monitor:handle_call |
[+] 4d7ecdfec MB-55578: Simplify kv_stats_monitor refresh mechanism |
[+] 7145e80dc MB-58168: Only get statuses from ns_memcached once in ns_heart |
[+] 9604df586 MB-55578: Only get statuses from ns_memcached once in kv_monitor |
|
In these three tests were setting autofailover timeout to 1sec
Build | Failure detection time (sec) | Job |
---|---|---|
7.6.0-1327 | 9.6 | http://perf.jenkins.couchbase.com/job/hestia/9294/console |
7.6.0-1350 | 7.0 | http://perf.jenkins.couchbase.com/job/hestia/9299/console |
7.6.0-1352 | 130.8 | http://perf.jenkins.couchbase.com/job/hestia/9301/console |