Bug #58288
closedquincy: mon: pg_num_check() according to crush rule
0%
Description
Corresponding BZ: https://bugzilla.redhat.com/show_bug.cgi?id=2153654
Introduced here in Q: https://github.com/ceph/ceph/pull/39062
An attempt to fix (not part of Q): https://github.com/ceph/ceph/pull/44430
reverting as an intermediate step before fixing in Q and main: https://github.com/ceph/ceph/pull/49465
Updated by Matan Breizman over 1 year ago
- Subject changed from pg_num_check underflow to mon: pg_num_check underflow
Updated by Radoslaw Zarzynski over 1 year ago
- Status changed from New to In Progress
- Assignee set to Matan Breizman
Just updating the tracker's state to fit the reality.
Updated by Matan Breizman over 1 year ago
- Subject changed from mon: pg_num_check underflow to quincy: mon: pg_num_check() according to crush rule
After the revert is merged (https://github.com/ceph/ceph/pull/49465),
pg_num_check() will return to not taking the crush rule into account when calculating projected pg num and osd num.
We prefer pg_num_check() to work inaccurately rather than resulting in an underflow error (as reported in the BZ attached).
The fix for correct pg num count will be backported to Q after the revert is merged (main PR for the fix: https://github.com/ceph/ceph/pull/49507).
This tracker is opened to address the (future) fix* to pg_num_check() in Q.
*Taking crush rule into account and skipping the check when decreasing the pool size.
Updated by Matan Breizman over 1 year ago
- Related to Bug #57105: quincy: ceph osd pool set <pool> size math error added
Updated by Matan Breizman over 1 year ago
- Related to Bug #47062: The pg_num check when creating the pool may not be accurate enough added
Updated by Matan Breizman about 1 year ago
- Status changed from In Progress to Fix Under Review
- Pull request ID set to 50327
Revert is merged https://github.com/ceph/ceph/pull/49465.
PR#50327 pushed as the actual fix.
Updated by Matan Breizman 10 months ago
- Status changed from Fix Under Review to Resolved