Actions
Bug #55568
openmgr/dashboard: CephPGImbalance alert inaccuracies
% Done:
0%
Source:
Tags:
monitoring, alerts backport_processed
Backport:
quincy, pacific, reef
Regression:
No
Severity:
3 - minor
Reviewed:
Description
Description of problem¶
The CephPGImbalance might be inaccurate in some specific environments. The CephPGImbalance alert is going to be raised, if the amount of PGs of an OSD is 30% less than the average amount of PGs of all OSDs. The amount of PGs of an OSD is normally related to e.g. number of OSDs and size or class of device. For a well-functioning cluster the Balancer takes care of balancing the PGs.
In some cases e.g. if the cluster consists of a lot of bigger HDDs and some smaller SSDs, the alert might be raised. Although the configuration and balance of the PGs is handled correctly.
As a workaround the alert can be silenced.
There are two possible solution, currently:
- Just remove the CephPGImbalance alert
- Replace the CephPGImbalance alert by e.g. an alert based on the Balancer score
Environment¶
ceph version
string:- Platform (OS/distro/release):
- Cluster details (nodes, monitors, OSDs):
- Did it happen on a stable environment or after a migration/upgrade?:
- Browser used (e.g.:
Version 86.0.4240.198 (Official Build) (64-bit)
):
How reproducible¶
Steps:
- ...
Actual results¶
Please add logs and/or screenshots
Expected results¶
here
Additional info¶
here
Actions