Actions
Bug #44791
closedfs/ceph: support cross-quota-tree renames
% Done:
0%
Source:
Tags:
Backport:
Regression:
No
Severity:
3 - minor
Reviewed:
Affected Versions:
ceph-qa-suite:
Crash signature (v1):
Crash signature (v2):
Description
In ceph_rename, we have this, which is probably what you're hitting:
/* don't allow cross-quota renames */
if ((old_dir != new_dir) &&
(!ceph_quota_is_same_realm(old_dir, new_dir)))
return -EXDEV;
The userspace client only returns EXDEV (or EDQUOT for renaming directories) when the rstats on the moving file/tree would put the target directory over its quota, which seems like a better behavior.
Files
Updated by Luis Henriques about 4 years ago
This behaviour was added to the kernel to mimic userspace commit bbfeaaea53f1 ("client: Forbid moving files across quota trees").
In the meantime, the userspace client changed its behaviour with commit b8954e5734b3 ("client: optimize rename operation under different quota root"). Unfortunately, this change has never been ported to the kernel client.
Actions
#2
Updated by Luis Henriques about 4 years ago
- File 0001-ceph-normalize-delta-parameter-usage-in-check_quota_.patch 0001-ceph-normalize-delta-parameter-usage-in-check_quota_.patch added
- File 0002-ceph-allow-rename-operation-under-different-quota-re.patch 0002-ceph-allow-rename-operation-under-different-quota-re.patch added
I'm attaching two patches that should implement in the kernel client the fuse client behaviour. They should hit the ceph-devel mailing-list soon.
Updated by Jeff Layton almost 4 years ago
- Status changed from New to Resolved
- Assignee set to Jeff Layton
Merged and should make v5.8.
Actions