Bug #62165
closed
[pg-autoscaler] Peformance issue with the autoscaler when we have ~32768 PGs
Added by Kamoltat (Junior) Sirivadhna 10 months ago.
Updated about 2 months ago.
Category:
pg_autoscaler module
Backport:
reef, quincy, pacific
Description
We have observed a performance issue regarding the autoscaler while scale testing using reef. The cpu peaking to up to a core every minute
Files
- Pull request ID set to 52633
The problem is due to how we have been extravagantly using osdmap get calls in the autoscaler module which becomes very costly when there is a lot of PGs.
Also attaching PR id.
- Subject changed from [pg-autoscaler] Peformance issue with 32768 taking up 1 whole cpu core to [pg-autoscaler] Peformance issue with the autoscaler when we have ~32768 PGs
- Project changed from Ceph to mgr
- Category set to pg_autoscaler module
- Status changed from New to Fix Under Review
- Status changed from Fix Under Review to Pending Backport
- Backport set to reef, quincy, pacific
- Copied to Backport #62885: reef: [pg-autoscaler] Peformance issue with the autoscaler when we have ~32768 PGs added
- Copied to Backport #62886: quincy: [pg-autoscaler] Peformance issue with the autoscaler when we have ~32768 PGs added
- Copied to Backport #62887: pacific: [pg-autoscaler] Peformance issue with the autoscaler when we have ~32768 PGs added
- Tags set to backport_processed
- Status changed from Pending Backport to Resolved
- % Done changed from 0 to 100
Also available in: Atom
PDF