Rook
v1.20.10Storage & DataNov 5, 2026
A maintenance release focused on redacting sensitive output, dependency updates, and correctness fixes. It also changes behavior and defaults for safer, more predictable operation, including object bucket relinking and Ceph global ID reclaim.
Check if affected (2)
breaking
OBCrelinking of another user's bucket is refusedApplies if you use
OBCto relink a bucket owned by another user.The
OBCrelink operation is refused when the bucket is owned by another user.breakingInsecure global ID reclaim is disabled
Applies if
cephis running.For
ceph, insecure global ID reclaim is disabled instead of muting the warning.
All 14 other recorded changesvalue changes 12 · additions 1 · fixes 1
value changes (12)
- - object: Log only the rgw zone fields rook reads (#18662, @jhoblitt)
- - core: Redact secret flag values from reported commands (#18665, @jhoblitt)
- - core: Shell-quote command lines in logs and sh -c strings (#18099, @jhoblitt)
- - security: Drop radosgw-admin output from object store errors (#18669, @jhoblitt)
- - cephfsmirror: Stop embedding the peer bootstrap token in an error (#18661, @jhoblitt)
- - security: Drop command output from unmarshal error text (#18335, @jhoblitt)
- - osd: Redact the lockbox secret from lvm zap output (#18670, @jhoblitt)
- - object: Redact credentials in CephBucketTopic endpoint logs (#18671, @jhoblitt)
- - core: Bump go-jose, swag and x/net to fix the snyk security scan (#18676, @jhoblitt)
- - mgr: Compare prometheus scrape interval numerically (#18583, @cvas22)
- - object: Emit rgwCommandFlags in a deterministic order (#18592, @travisn)
- - osd: Avoid empty label selector when all OSDs in batch are skipped (#18564, @trukhinyuri)
additions (1)
- - cluster: Allow configuring the Prometheus scrape timeout (#18424, @harshilv17)
fixes (1)
- - clusterdisruption: Fix minimum failure domain detection for region (#18577, @NUAAwqy)
Add Rook to your stack
A weekly email arrives when a release needs action. Like the breaking changes in this release.