Repository navigation

Issues

New issue

🚨 Ceph CVE notice and Rook advisory
#18203 · BlaineEXE opened on Aug 19, 2026
4
OOM-Kills of standby MGR after Upgrade to rook v1.20.1 / ceph v20.2.2
#17786 · bitfisher opened on Jun 19, 2026
23
Tentacle v20.2.3 on Rook: permanent mgr RECENT_MGR_MODULE_CRASH — prometheus get_hardware_metrics() → node_proxy_fullreport() NotImplementedError
#18124 · qveensi opened on Aug 10, 2026
25

In Multisite setups the system user credential doesn't rotate when secret updated
< 8000 /div>
rook/rook#18738 · opened on Nov 6, 2026
clusterdisruption: noout and blocking OSD PDBs never cleared after the node returns (stale cache read of the deleted default PDB, no OSD watch, osdMaintenanceTimeout only inside reconcile)
rook/rook#18687 · opened on Nov 4, 2026
Require either all pools or no pools to set the deviceClass
rook/rook#18679 · opened on Nov 3, 2026
Update docs generation to replace deprecated mkdocs
rook/rook#18675 · opened on Nov 3, 2026
6
Operator exits with installCsiOperator: false: rados namespace and subvolume group controllers watch ClientProfile without the CRD
rook/rook#18654 · opened on Nov 2, 2026
Ceph Vampire: Add support for 2AZ+arbiter stretched EC and replica pools
rook/rook#18646 · opened on Oct 29, 2026
5
CephObjectStore: rgwConfig / rgwConfigFromSecret keys are not removed from the mon config store when deleted from the CR
rook/rook#18645 · opened on Oct 29, 2026
2
Verify the daemon key rotation behavior during cluster upgrade when one or more OSDs are down.
rook/rook#18638 · opened on Oct 27, 2026
key-rotation: Skip rotating keys if OSD, MGR, MDS and MON daemons are not upgraded to desired version.
rook/rook#18637 · opened on Oct 27, 2026
mon PDB maxUnavailable lets a second mon be evicted during parallel node drains
rook/rook#18625 · opened on Oct 23, 2026
external: detach fails for clusters imported with --restricted-auth-permission
rook/rook#18619 · opened on Oct 22, 2026
v1.21: CSI driver crashes on x86-64-v2 CPUs after upgrade — dropped cephcsi default pulls v3.18.1 (Rocky Linux 10, needs x86-64-v3)
rook/rook#18607 · opened on Oct 20, 2026
2
Multisite Setup Requires manual system user creation
rook/rook#18606 · opened on Oct 20, 2026
3
CSI addons fails to create CSIAddonsNode: no matches for csiaddons.openshift.io/v1alpha1
Bug
rook/rook#18591 · opened on Oct 19, 2026
5
external: --restricted-auth-permission ignores --rbd-metadata-ec-pool-name, producing RBD keys that cannot provision
rook/rook#18586 · opened on Oct 19, 2026
mon: a mon that stops does not leave the quorum first; OSDs that stop at the same time block writes for about 10 s
rook/rook#18581 · opened on Oct 16, 2026
2
external: --restricted-auth-permission does not restrict client.healthchecker (osd profile rbd-read-only on all pools, mds allow *)
rook/rook#18575 · opened on Oct 16, 2026
RBD Regional-DR pool-ID mapping (cluster-mapping.json) is never delivered to CSI pods in Rook-managed clusters
rook/rook#18556 · opened on Oct 13, 2026
8
key rotation can take place when not all OSDs are upgraded if nodes are cordoned
rook/rook#18537 · opened on Oct 9, 2026
4
CephBlockPool pg_num_min / pg_num parameters fail depending on the order they are applied
rook/rook#18529 · opened on Oct 8, 2026
1
helm: aggregate Ceph custom resources into the builtin view, edit and admin ClusterRoles
rook/rook#18519 · opened on Oct 7, 2026
CephNodeNetworkPacketErrors for unrelated node
rook/rook#18496 · opened on Oct 4, 2026
4
OSD blocking PDBs are never removed after Ceph heals around a cordoned node
rook/rook#18492 · opened on Oct 2, 2026
2
Raw-mode OSD activate deadlocks on nodes with hung krbd devices: ceph-volume raw activate runs lvs (Ceph >= v20.2.3), rbd LVM filter is only applied in lvm mode
rook/rook#18483 · opened on Oct 1, 2026
5
OBC user creation fails when CephObjectStore is an RGW multisite zone not in the master zonegroup
rook/rook#18479 · opened on Sep 30, 2026
1

Navigation Menu

Repository navigation

Issues

Issues

In Multisite setups the system user credential doesn't rotate when secret updated

clusterdisruption: noout and blocking OSD PDBs never cleared after the node returns (stale cache read of the deleted default PDB, no OSD watch, osdMaintenanceTimeout only inside reconcile)

Require either all pools or no pools to set the deviceClass

Update docs generation to replace deprecated mkdocs

Operator exits with `installCsiOperator: false`: rados namespace and subvolume group controllers watch ClientProfile without the CRD

Ceph Vampire: Add support for 2AZ+arbiter stretched EC and replica pools

CephObjectStore: rgwConfig / rgwConfigFromSecret keys are not removed from the mon config store when deleted from the CR

Verify the daemon key rotation behavior during cluster upgrade when one or more OSDs are down.

key-rotation: Skip rotating keys if OSD, MGR, MDS and MON daemons are not upgraded to desired version.

mon PDB maxUnavailable lets a second mon be evicted during parallel node drains

external: detach fails for clusters imported with --restricted-auth-permission

v1.21: CSI driver crashes on x86-64-v2 CPUs after upgrade — dropped cephcsi default pulls v3.18.1 (Rocky Linux 10, needs x86-64-v3)

Multisite Setup Requires manual system user creation

CSI addons fails to create CSIAddonsNode: no matches for csiaddons.openshift.io/v1alpha1

external: --restricted-auth-permission ignores --rbd-metadata-ec-pool-name, producing RBD keys that cannot provision

mon: a mon that stops does not leave the quorum first; OSDs that stop at the same time block writes for about 10 s

external: --restricted-auth-permission does not restrict client.healthchecker (osd profile rbd-read-only on all pools, mds allow *)

RBD Regional-DR pool-ID mapping (cluster-mapping.json) is never delivered to CSI pods in Rook-managed clusters

key rotation can take place when not all OSDs are upgraded if nodes are cordoned

CephBlockPool pg_num_min / pg_num parameters fail depending on the order they are applied

helm: aggregate Ceph custom resources into the builtin view, edit and admin ClusterRoles

CephNodeNetworkPacketErrors for unrelated node

OSD blocking PDBs are never removed after Ceph heals around a cordoned node

Raw-mode OSD activate deadlocks on nodes with hung krbd devices: ceph-volume raw activate runs lvs (Ceph >= v20.2.3), rbd LVM filter is only applied in lvm mode

OBC user creation fails when CephObjectStore is an RGW multisite zone not in the master zonegroup