Bug #66960
closedmon: fix `fs set down` to adjust max_mds only when cluster is not down
0%
Description
Applying `fs set down false` on an up cluster will set the cluster's max_mds to 1, regardless of the cluster's current max_mds.
`fs set down false` should only change max_mds when the cluster was set to down. Otherwise, the cluster should remain unchanged.
Files
Updated by Radoslaw Zarzynski over 1 year ago
- Project changed from RADOS to CephFS
Updated by Venky Shankar over 1 year ago
- Category set to Administration/Usability
- Status changed from New to Fix Under Review
- Assignee set to chung feng zheng
- Target version set to v20.0.0
- Source set to Community (user)
- Backport set to quincy,reef,squid
Updated by Rishabh Dave over 1 year ago
This bug can't be reproduced. chungfengz Did you see this bug on one of the older branches? What steps you followed to reproduce it? This bug can't be reproduced on the @main branch.
Updated by chung feng zheng over 1 year ago
- File cephfs.png cephfs.png added
Rishabh Dave wrote in #note-3:
This bug can't be reproduced.
chungfengz Did you see this bug on one of the older branches? What steps you followed to reproduce it? This bug can't be reproduced on the @mainbranch.
To reproduce this bug, follow these steps:
1. Set up a CephFS cluster with multiple MDS (Metadata Servers), ensuring max_mds is set to a value greater than 1.
2. Use the command ceph fs set <cephfs> down false. This will cause max_mds to be reset to 1.
Note: This bug only manifests in a multi-MDS cluster or when max_mds is greater than 1.
Updated by Rishabh Dave over 1 year ago
chung feng zheng wrote in #note-4:
Rishabh Dave wrote in #note-3:
This bug can't be reproduced.
chungfengz Did you see this bug on one of the older branches? What steps you followed to reproduce it? This bug can't be reproduced on the @mainbranch.To reproduce this bug, follow these steps:
1. Set up a CephFS cluster with multiple MDS (Metadata Servers), ensuring max_mds is set to a value greater than 1.
2. Use the command ceph fs set <cephfs> down false. This will cause max_mds to be reset to 1.Note: This bug only manifests in a multi-MDS cluster or when max_mds is greater than 1.
I see. The bug can only be reproduced in a case when down is being to false without ever setting it to true. And, ofcourse, max_mds > 1 in this case. The PR is fine in that case.
Updated by Patrick Donnelly over 1 year ago
- Status changed from Fix Under Review to Pending Backport
- Backport changed from quincy,reef,squid to squid,reef,quincy
Updated by Upkeep Bot over 1 year ago
- Copied to Backport #67373: quincy: mon: fix `fs set down` to adjust max_mds only when cluster is not down added
Updated by Upkeep Bot over 1 year ago
- Copied to Backport #67374: squid: mon: fix `fs set down` to adjust max_mds only when cluster is not down added
Updated by Upkeep Bot over 1 year ago
- Copied to Backport #67375: reef: mon: fix `fs set down` to adjust max_mds only when cluster is not down added
Updated by Upkeep Bot over 1 year ago
- Tags (freeform) set to backport_processed
Updated by Upkeep Bot 9 months ago
- Merge Commit set to 95dffdcbbd78689670489ee712bbab0c024fcabd
- Fixed In set to v19.3.0-3997-g95dffdcbbd7
- Upkeep Timestamp set to 2025-07-08T22:37:58+00:00
Updated by Upkeep Bot 8 months ago
- Fixed In changed from v19.3.0-3997-g95dffdcbbd7 to v19.3.0-3997-g95dffdcbbd78
- Upkeep Timestamp changed from 2025-07-08T22:37:58+00:00 to 2025-07-14T15:46:32+00:00
Updated by Upkeep Bot 8 months ago
- Fixed In changed from v19.3.0-3997-g95dffdcbbd78 to v19.3.0-3997-g95dffdcbbd
- Upkeep Timestamp changed from 2025-07-14T15:46:32+00:00 to 2025-07-14T21:10:37+00:00
Updated by Upkeep Bot 7 months ago
- Status changed from Pending Backport to Resolved
- Upkeep Timestamp changed from 2025-07-14T21:10:37+00:00 to 2025-08-25T19:42:34+00:00
Updated by Upkeep Bot 5 months ago
- Released In set to v20.2.0~2291
- Upkeep Timestamp changed from 2025-08-25T19:42:34+00:00 to 2025-11-01T01:37:21+00:00