Actions
Bug #38604
closedmon logs not getting reopened after rotation
Status:
Resolved
Priority:
Immediate
Assignee:
-
Category:
-
Target version:
-
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Component(RADOS):
Pull request ID:
Crash signature (v1):
Crash signature (v2):
Description
root@mira055:~/ceph-deploy# ls /var/log/ceph/*mon*log ls: cannot access '/var/log/ceph/*mon*log': No such file or directory root@mira055:~/ceph-deploy# ls /var/log/ceph/*mon*log* /var/log/ceph/ceph-mon.mira055.log.10.gz /var/log/ceph/ceph-mon.mira055.log.11.gz /var/log/ceph/ceph-mon.mira055.log.12.gz /var/log/ceph/ceph-mon.mira055.log.1.gz /var/log/ceph/ceph-mon.mira055.log.2.gz /var/log/ceph/ceph-mon.mira055.log.3.gz /var/log/ceph/ceph-mon.mira055.log.4.gz /var/log/ceph/ceph-mon.mira055.log.5.gz /var/log/ceph/ceph-mon.mira055.log.6.gz /var/log/ceph/ceph-mon.mira055.log.7.gz /var/log/ceph/ceph-mon.mira055.log.9.gz root@mira055:~/ceph-deploy# ls /var/log/ceph/*osd*log /var/log/ceph/ceph-osd.0.log /var/log/ceph/ceph-osd.2.log /var/log/ceph/ceph-osd.13.log /var/log/ceph/ceph-osd.3.log /var/log/ceph/ceph-osd.23.log /var/log/ceph/ceph-osd.73.log /var/log/ceph/ceph-osd.26.log /var/log/ceph/ceph-osd.admin.log
a kill -1 reopens it find. not sure why teh automatic rotation isn't working
seems to be the case with the lab cluster mons. reesi001 i manually HUPed, the others are untouched.
Updated by Sage Weil about 5 years ago
this appears to be because of /proc/$pid/stat. before,
598968 (exe) S 1 598968 598968 0 -1 1077936128 3642186 636 181 0 161819 54802 14 28 20 0 25 0 3429284895 1389723648 202301 18446744073709551615 4194304 13828103 140722758060896 140722758051408 140239223620448 0 4096 4096 1098925291 0 0 0 17 7 0 0 947 0 0 15928352 16043120 34836480 140722758065813 140722758065894 140722758065894 140722758066153 0
after restarting ceph-mon,
2246944 (ceph-mon) S 1 2246944 2246944 0 -1 1077936384 15888 0 86 0 208 66 0 0 20 0 25 0 3439136158 521064448 19056 18446744073709551615 4194304 13828103 140732176569424 140732176559936 139678339625824 0 4096 4096 1098925291 0 0 0 17 6 0 0 8 0 0 15928352 16043120 17850368 140732176572165 140732176572246 140732176572246 140732176572390 0
that second word in parens is what killall is matching.
i think it is "(exe)" instead of "(ceph-mon)" because ceph-mon did respawn when msgr2 was enabled. If so, then ceph-mds and ceph-mgr probably have similar problems, because they also (regularly) respawn.
Updated by Sage Weil about 5 years ago
- Status changed from 12 to Fix Under Review
aha, ceph-mgr and ceph-mds expliiclty set the thread name on startup.
Updated by Sage Weil about 5 years ago
- Status changed from Fix Under Review to Resolved
Actions