Hi there, I have seen this behaviour due to an issue in the design of access log buffer flushing. During a buffer flush all other threads are delayed, which can cause this spike. You can confirm this by changing your access log buffer size up or down. Sadly this problem is not simply fixed without a complete rewrite of the logging subsystem - which I would love to do, but I am not employed to work on 389 at this time so I lack the time. Hope that helps. On Wed, 2018-09-05 at 21:14 +0000, Ghiurea, Isabella wrote: > > Hello Gurus, > looking for an answer to the following performance behavior > my env: 389-ds-base-1.3.5.15-1.fc24.x86_64 in multimaster fractional > replication > > running rsearch for 5 min with 1 thread seeing spikes for a basic > read using index uid > And running with 10 threads same search the avg ms/ops performance > are much better with no major spike/burst > Any explanation much appreciate it > > see bellow for 1 thread and the spike/burst > T 300 -t 1 > rsearch: 1 threads launched. > T1 min= 0ms, max= 5ms, count = 54710 > T1 min= 0ms, max= 42ms, count = 64930 > T1 min= 0ms, max= 2ms, count = 65174 > T1 min= 0ms, max= 2ms, count = 65110 > T1 min= 0ms, max= 44ms, count = 64966 > T1 min= 0ms, max= 1ms, count = 65101 > T1 min= 0ms, max= 22ms, count = 65056 > T1 min= 0ms, max= 32ms, count = 64981 > T1 min= 0ms, max= 1ms, count = 65145 > T1 min= 0ms, max= 1ms, count = 65223 > T1 min= 0ms, max= 27ms, count = 65015 > T1 min= 0ms, max= 1ms, count = 65182 > T1 min= 0ms, max= 3ms, count = 65213 > T1 min= 0ms, max= 23ms, count = 64760 > T1 min= 0ms, max= 2ms, count = 64214 > T1 min= 0ms, max= 3ms, count = 52279 > T1 min= 0ms, max= 11ms, count = 64914 > T1 min= 0ms, max= 1ms, count = 65118 > T1 min= 0ms, max= 5ms, count = 64852 > T1 min= 0ms, max= 91ms, count = 64180 > T1 min= 0ms, max= 4ms, count = 64746 > T1 min= 0ms, max= 1ms, count = 65080 > T1 min= 0ms, max= 12ms, count = 65110 > T1 min= 0ms, max= 702ms, count = 59243 > T1 min= 0ms, max= 1ms, count = 65082 > T1 min= 0ms, max= 89ms, count = 64331 > T1 min= 0ms, max= 23ms, count = 64647 > T1 min= 0ms, max= 5ms, count = 64818 > T1 min= 0ms, max= 55ms, count = 64374 > T1 min= 0ms, max= 8ms, count = 64713 > T1 min= 0ms, max= 8ms, count = 64713 > 300 sec >= 300 > Final Average rate: 6394.22/sec = 0.1564msec/op, total: 64713 > > And final avg rate for 10 threads, no significant spike/burst for > this num of threads > > 20180905 14:07:23 - Rate: 16962.10/thr (16962.10/sec = 0.0590ms/op), > total:169621 (10 thr) > 300 sec >= 300 > Final Average rate: 17420.40/sec = 0.0574msec/op, total:169621 > > > _______________________________________________ > 389-users mailing list -- 389-users@xxxxxxxxxxxxxxxxxxxxxxx > To unsubscribe send an email to > 389-users-leave@xxxxxxxxxxxxxxxxxxxxxxx > Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html > List Guidelines: > https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: > https://lists.fedoraproject.org/archives/list/389-users@xxxxxxxxxxxxxxxxxxxxxxx -- Sincerely, William _______________________________________________ 389-users mailing list -- 389-users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to 389-users-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/389-users@xxxxxxxxxxxxxxxxxxxxxxx
