Ceph Logs Not Rotated on MOS 8.0
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Mirantis OpenStack |
Confirmed
|
High
|
MOS Maintenance |
Bug Description
Detailed Bug description:
When deploying a node with MOS 8.0 and selecting ceph during environment configuration, some ceph logs on the controller nodes are not rotated. There doesn't appear to be any logrotate configuration for ceph in /etc/logrotate.d/. This can lead to issues as some logs grow rather large (I currently have a multi GB log for ceph-client.
Steps To Reproduce:
1) MOS 8.0 Fuel
2) Deploy 1 controller, 2 compute+storage (storage configured to use ceph - I have two enviroments that use ceph. First environment uses it for everything, second environment everything except ephemeral)
3) Log into controller, notice /etc/logrotate.
Expected Results:
/etc/logrotate.
Actual Results:
/etc/logrotate.
Workaround:
Manually create /etc/logrotate.
Impact:
/var/log fills up, pacemaker cannot find any machine to run resources because #health_check fails on single controller when /var/log is almost full (>95% in my case, I don't know when #health_check fails)
Changed in mos: | |
status: | New → Confirmed |
importance: | Undecided → High |
assignee: | nobody → MOS Maintenance (mos-maintenance) |
milestone: | none → 8.0-updates |
tags: | added: area-ceph |
I think this also applies to /var/log/radosgw. It looks like that is where ceph-client. radosgw. gateway. log lands, not in /var/log/ceph/. Both directories probably need rotate setup for log files.