Actions
Bug #43975
closedSlow Requests/OP's types not getting logged
% Done:
0%
Source:
Tags:
Backport:
nautilus
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
MonClient
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
- From ceph.log
2020-01-30 05:00:34.367139 mon.node1 (mon.0) 53851 : cluster [WRN] Health check update: Degraded data redundancy: 5916359/1069099881 objects degraded (0.553%), 201 pgs degraded (PG_DEGRADED) 2020-01-30 05:00:39.368143 mon.node1 (mon.0) 53853 : cluster [WRN] Health check update: Degraded data redundancy: 5916361/1069100427 objects degraded (0.553%), 201 pgs degraded (PG_DEGRADED) 2020-01-30 05:00:41.100123 mon.node1 (mon.0) 53854 : cluster [WRN] Health check failed: 668 slow ops, oldest one blocked for 207 sec, daemons [osd,0,osd,100,osd,102,osd,106,osd,108,osd,109,osd,110,osd,112,osd,119,osd,120]... have slow ops. (SLOW_OPS) 2020-01-30 05:00:32.949857 mgr.node1 (mgr.14052) 108028 : cluster [DBG] pgmap v109693: 4768 pgs: 1 active+recovery_wait, 8 active+undersized, 151 active+undersized+degraded, 50 active+recovery_wait+degraded, 6 active+clean+scrubbing, 4552 active+clean; 2.4 TiB data, 75 TiB used, 285 TiB / 361 TiB avail; 9.0 MiB/s rd, 1.2 MiB/s wr, 4.19k op/s; 5916359/1069099881 objects degraded (0.553%); 11 KiB/s, 0 objects/s recovering
- Earlier in the luminous, we used to log slow requests/ops type, for example -
reached pg queued for pg op applied waiting for sub ops
Files
Actions