Back to bug 1375538

Who When What Removed Added
Martin Bukatovic 2016-09-13 10:51:51 UTC See Also https://bugzilla.redhat.com/show_bug.cgi?id=1366577
Martin Bukatovic 2016-09-13 10:52:23 UTC CC linuxkidd
Jeff Applewhite 2016-09-14 12:47:12 UTC CC japplewh
Shubhendu Tripathi 2016-09-22 07:18:05 UTC CC julim, shtripat
Flags needinfo?(julim) needinfo?(linuxkidd)
Michael J. Kidd 2016-09-22 16:59:17 UTC Flags needinfo?(julim) needinfo?(linuxkidd)
Shubhendu Tripathi 2016-09-23 06:51:39 UTC Flags needinfo?(julim)
Shubhendu Tripathi 2016-09-30 05:51:53 UTC Status NEW POST
Link ID Gerrithub.io 295614
Assignee nthomas shtripat
Shubhendu Tripathi 2016-09-30 08:27:08 UTC Status POST MODIFIED
Fixed In Version rhscon-ceph-0.0.43-1.el7scon.x86_64, rhscon-ui-0.0.58-1.el7scon.noarch
Sharmilla Abhilash 2016-10-03 10:50:31 UTC CC vsarmila
Blocks 1357777
Shubhendu Tripathi 2016-10-03 11:12:43 UTC Status MODIFIED ON_QA
Martin Bukatovic 2016-10-03 12:11:32 UTC QA Contact sds-qe-bugs mbukatov
Martin Bukatovic 2016-10-03 16:21:16 UTC Status ON_QA ASSIGNED
Martin Bukatovic 2016-10-03 16:32:39 UTC Flags needinfo?(linuxkidd)
Michael J. Kidd 2016-10-03 20:12:03 UTC Flags needinfo?(julim) needinfo?(linuxkidd)
Karnan 2016-10-04 06:24:11 UTC CC kchidamb
Flags needinfo?(linuxkidd)
Karnan 2016-10-04 06:25:23 UTC Flags needinfo?(julim)
Michael J. Kidd 2016-10-04 20:37:54 UTC Flags needinfo?(linuxkidd) needinfo?(julim)
Karnan 2016-10-06 09:49:55 UTC Status ASSIGNED MODIFIED
Fixed In Version rhscon-ceph-0.0.43-1.el7scon.x86_64, rhscon-ui-0.0.58-1.el7scon.noarch rhscon-ceph-0.0.43-1.el7scon.x86_64, rhscon-ui-0.0.60-1.el7scon.noarch
Martin Bukatovic 2016-10-06 15:08:22 UTC Status MODIFIED VERIFIED
Nishanth Thomas 2016-10-14 10:44:45 UTC CC nthomas
Doc Type If docs needed, set a value Bug Fix
Shubhendu Tripathi 2016-10-14 14:00:41 UTC Doc Text Cause:
Earlier the PG no value used to be calculated using a prescribed logic within the backend during pool creation and for certain ranges like 0-5, 5-10, 10-50 OSDs the values used to be hardcoded.

Consequence:
Cluster used to go to WARN state due to more and more no of PGs created during each pool creation.

Fix:
As a suggestion from ceph team, disabled the automatic calculation logic for the PGs while pool creation. Rather ask admin to provide the PG value referring the PG calculation pool from ceph.

Result:
The cluster would no more go to WARN state as long as correct PG value if calculated using PG calc tool and passed in request to backend.
Rakesh 2016-10-17 12:17:02 UTC CC rghatvis
Doc Text Cause:
Earlier the PG no value used to be calculated using a prescribed logic within the backend during pool creation and for certain ranges like 0-5, 5-10, 10-50 OSDs the values used to be hardcoded.

Consequence:
Cluster used to go to WARN state due to more and more no of PGs created during each pool creation.

Fix:
As a suggestion from ceph team, disabled the automatic calculation logic for the PGs while pool creation. Rather ask admin to provide the PG value referring the PG calculation pool from ceph.

Result:
The cluster would no more go to WARN state as long as correct PG value if calculated using PG calc tool and passed in request to backend.
Previously, the automatic PG calculation logic caused problems as it calculated on per pool basis instead of calculating on a cluster level based on the number of OSDs in the cluster and PGs should be shared across the pools in the cluster. This incorrect PG calculation issued cluster health warning due to large number of PGs being created during each pool creation.

With this update, the automatic calculation of PGs is disabled. The administrator needs to manually provide the PG values per OSD by using the PG calculator tool from Ceph to ensure the cluster remains in a healthy state.
Flags needinfo?(shtripat)
Shubhendu Tripathi 2016-10-17 12:20:51 UTC Flags needinfo?(shtripat)
errata-xmlrpc 2016-10-19 15:22:28 UTC Status VERIFIED CLOSED
Resolution --- ERRATA
Last Closed 2016-10-19 11:22:28 UTC

Back to bug 1375538