When the operator starts, there is no need to handle the
node watcher immediately while the controller runtime
cache is being initialized. The cluster will anyway be
fully reconciled and create new OSDs if needed. The node
watcher is only intended for triggering a reconcile when
a node is added or updated at a later time after a sync
would anyway be triggered.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 7803fac10d)
adding check if operator is installed via helm with
existing cluster chart check so that if some users
install only operator chart and not the cluster chart
we are still able to transfer the ownership of old
yaml based csi resource to new helm based csi.
Signed-off-by: subhamkrai <srai@redhat.com>
When RGW and MDS intances are reduce to 1, the older RGW and MDS PDB
instance (with maxUnavailabe=1) are still available. This blocks node
drain when the node, on which the single RGW and MDS instances are
running, is drained.
This PR deletes any stale MDS and RGW PDB when the only 1 MDS and 1 RGW
instances are available.
Signed-off-by: Santosh <sapillai@redhat.com>
(cherry picked from commit 3396198e9c)
Add support for the osd.rook.io/device-class node label to assign
CRUSH device classes to all OSDs on a node.
The operator reads the label during OSD provisioning and
deployment reconciliation. If CR-level deviceClass conflicts
with the node label, the operator will skip reconciling the
node and logs an error.
Resolves: #17131
Signed-off-by: Yuxiang He <yuxiang@dropbox.com>
(cherry picked from commit d9357dbfdc)
applyUpgradeOSDFunctionality() may run before all OSDs have registered
their new version, leaving require-osd-release unset and the cluster in
HEALTH_WARN.
Check for a single converged OSD version on each health-monitor tick and
set require-osd-release when it changes. The release name is cached to
avoid redundant ceph commands
Signed-off-by: Sunnatillo <sunnat.samadov@est.tech>
(cherry picked from commit f6fa25e844)
User RaceFreeFakeWatcher isntead of FakeWatcher in TestAddRemoveNode
function. The test runs c.Start() in a goroutine whch stops the watcher
when done. If mockNodeOrchestrationCompeltion sends a watch ever after,
that, FakeWatcher panics with "send on a closed channel".
RaceFreeFakeWatcher safely ignores the sends on a stopped channel.
Signed-off-by: Santosh <sapillai@redhat.com>
(cherry picked from commit 6781d2e58d)
Fixes#16834 and #17384. Per-OSD CR deviceClass was ignored in
osd prepare job when reporting the osd config back to the operator,
and in the operator's reconcile override.
Fixed by resolving the per-device class from the CR devices list
in GetCephVolumeRawOSDs and in makeDeployment.
Signed-off-by: Artem Torubarov <artem.torubarov@sap.com>
(cherry picked from commit b41d63788d)
Clean up duplicated words and a typo in object storage docs, comments, and user-facing messages.
Signed-off-by: immanuwell <pchpr.00@list.ru>
(cherry picked from commit c3a9f5fdd9)
Previously, shouldFailoverMonImmediately() was called only when a mon first went out of quorum. If the node gets deleted after the timeout starts, it does not get detected until the timeout expires. Moved the call outside the timeout initialization block so it runs on every health check iteration and detects node deletion immediately.
Fixes#15953
Signed-off-by: Prabhala Tara Aasrita <taraasrita@ibm.com>
(cherry picked from commit aeae61e5be)
Detect OSD encryption from the dmcrypt block path in `getOSDInfo` as a
fallback when the `encrypted` label is missing on pre-label deployments.
Include `osd.Encrypted` in the label check in `makeDeployment` so the
label reflects the actual OSD state on upgrade.
Fixes a bug where upgrading to a Rook version with the `encrypted` label
would set the label to `false` on already-encrypted OSDs, because the
label didn't exist yet and `getOSDInfo` defaulted to `false`.
Signed-off-by: Santosh <sapillai@redhat.com>
(cherry picked from commit fc9e3b54f3)
rook-ceph-exporter should use DNSPolicy ClusterFirstWithHostNet when
hostNetwork is enabled
Signed-off-by: Rohan Gupta <rohgupta@redhat.com>
(cherry picked from commit 17873d3af2)
Strip unnecessary leading and trailing newlines from osdLivenessProbeScript and livenessProbeScript raw string literals in spec.go
Signed-off-by: Santosh <sapillai@redhat.com>
(cherry picked from commit 925acb3e92)
We are moving the CSI management to the admin and currently, the
CSI resources are created at runtime so the cluster installed with
helm will face an error that resources are not managed by helm.
To avoid this error, we are adding the annotation at runtime so
that during upgrade users do not face issues.
Signed-off-by: subhamkrai <srai@redhat.com>
When an OSD is purged and reprovisioned, it comes back without the
correct CRUSH device class. NVMe devices configured with
deviceClass: nvme in the CR end up as ssd.
Root cause: the operator sets ROOK_OSD_CRUSH_DEVICE_CLASS on prepare
jobs but not on OSD deployments. During re-discovery
(GetCephVolumeRawOSDs), the daemon reads this env var, finds it
absent, and falls back to auto-detected deviceType which returns ssd
for NVMe.
Fix: add crushDeviceClassEnvVar(osd.DeviceClass) to the OSD
deployment spec in spec.go, alongside the existing
dataDeviceClassEnvVar. Existing deployments pick up the new env var
on operator reconciliation without manual intervention.
Fixes#17227
Signed-off-by: David Orman <ormandj@corenode.com>
(cherry picked from commit 1357d71177)
The mon PDB was previously only being adjusted
during mon failover, where the maxunavailable was set to 0
before failover, then reset to 1 after failover. This allowed
for a race condition where one mon had been drained and is
not fully back in quorum before the next mon is allowed to drain.
This is now prevented by disallowing drains until a mon
is fully back in quorum. Anytime a mon falls out of quorum,
the mon pdb will be adjusted to prevent further drains
from causing a mon quorum outage.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 80b50969b2)
Previously provisionerPriorityClassName config is applied to node plugin
but pluginPriorityClassName is applied to controller plugin. This causes
a misconfiguration since the default values for
provisionerPriorityClassName is system-cluster-critical and
pluginPriorityClassName is system-node-critical.
The default priority for node plugin ought to be system-node-critical so
we can leverage this information to kill the pods in this priority last,
allowing other pods that may mount RBD-backed volume to unmount cleanly
first, otherwise the shutdown sequence may hang due to failure
unmounting RBD-backed volume when the node plugin pod is already killed.
Signed-off-by: Sonny Sasaka <sonnysasaka@gmail.com>
(cherry picked from commit dff04494fe)
This parameter is written to .spec.cephFS.radosNamespace of the generated
ClientProfile.
The parameter is called 'CSIMetadataRadosNamespace' to clarify that it is used
for CSI metadata and unrelated to the actual CephFS (meta)data.
Context:
The ceph-csi CephFS plugin stores additional metadata related to PV(C)s
in RADOS objects in the metadata pool of the CephFS.
By default those objects are stored in the 'csi' RADOS namespace.
For a multi-tenant CephFS setup (that is used by multiple K8s Clusters/
ceph-csi drivers) segregating those additional objects into separate RADOS
namespaces is desireable so the clusters can not modify the metadata of other
clusters (enforced by appropriate ceph client capabilities).
'ceph-csi' implemented this via the `cephFS.radosNamespace` config option for
cluster entries in the `ceph-csi-config` ConfigMap with https://github.com/ceph/ceph-csi/pull/4661.
The `ceph-csi-operator` added support for that config entry by adding
'radosNamespace' to the 'CephFsConfigSpec' of the 'ClientProfile' with
https://github.com/ceph/ceph-csi-operator/pull/165.
Signed-off-by: Alexander Degenhart <degenhart@fim.uni-passau.de>
(cherry picked from commit 17b98da89e)
When osd versions are being detected, log more info
about the versions to understand better if the
require-osd-version release is being set as
expected.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit a80ca49855)
This prevents overwriting the user-specified capabilities
and correctly copies the Capabilities into UserCapabilities
after the AdminOpsClient calls.
Signed-off-by: hjk068 <hello.hyunjin@gmail.com>
(cherry picked from commit b113e37dae)
Allow users to configure custom labels on CephObjectStore RGW services
by adding a Labels field to the RGWServiceSpec, following the existing
pattern used for Annotations. This enables use cases such as service
mesh integration and monitoring discovery that require specific labels
on Kubernetes services.
Fixes: #17235
Signed-off-by: majiayu000 <1835304752@qq.com>
(cherry picked from commit ac6ec79865)
deleting an erasure coded CephBlockPool left the ec profile
orphaned. recreating with different settings failed because
the old profile could not be overridden.
- delete ec profile during CephBlockPool deletion
- disable --format json on ec profile set command
- fix ec profile name mismatch in CephObjectStore deletion
Signed-off-by: Oded Viner <oviner@redhat.com>
(cherry picked from commit b88648d10f)
ec pools skip the mirroring configuration path where
the status is set to ready, leaving them stuck in
progressing state forever. add status update to ready
in the else branch when mirroring is skipped.
Signed-off-by: Oded Viner <oviner@redhat.com>
(cherry picked from commit ba9fad77a0)
Clean up stale exporter deployments whose target node no longer exists.
This handles the case where the operator was down when a node was deleted,
so the deletion event was missed and the stale deployment was never removed.
Signed-off-by: Muhammad Adil Ghaffar <muhammad.adil.ghaffar@est.tech>
(cherry picked from commit 0b8ed65749)
When reconciling the crash collectors and exporters, if one cluster
has disabled them, they will be deleted, then the reconcile
will complete, thus skipping reconcile of the crash collectors
and exporters for the remaining clusters.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 71e1dbffa4)
Remove the exporter deployment when needed as best effort
in case the removal fails, there is no need to block other
reconcile.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 256a338c14)
Rook no longer supports earlier than Ceph v18, so remove the
obsolete version checking in the ceph exporter reconcile.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 733406b175)
the ec pool doesnt support mirroring from ceph end
so skip if the mirroring spec is set on the ec pools
Signed-off-by: parth-gr <partharora1010@gmail.com>
(cherry picked from commit 9f932d477e)
the ceph exporter port was hardcoded to 9926, preventing
co-location of exporters from different ceph clusters on
the same node. add a port field to cephexporterspec so
users can override it via spec.monitoring.exporter.port.
Signed-off-by: Oded Viner <oviner@redhat.com>
(cherry picked from commit c2a2fc4241)
The osd unit tests periodically fail with a timing issue
when canceling an orchestration. Attempt to stabilize
the test with a longer timeout before canceling the
test.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 96f4320a34)
Updated the following csi sidecars to their latest available versions:
- csi-attacher: v4.11.0
- csi-snapshotter: v8.5.0
- csi-resizer: v2.1.0
- csi-provisioner: v6.1.1
- csi-node-driver-registrar: v2.16.0
Signed-off-by: Praveen M <m.praveen@ibm.com>
(cherry picked from commit 6b55092abf)
Update `lockbox.keyring` rotation for encrypted OSDs to ensure the key
is not lost if the init container fails but continues assuming
(reasonably) that the key is likely still relevant.
Signed-off-by: Blaine Gardner <blaine.gardner@ibm.com>
(cherry picked from commit 931006fb91)
We should set OSDInfo.DeviceType if "device-type" label is
set on the corresponding Deployment. Otherwise, Rook
deletes this label on Update.
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
(cherry picked from commit dac71d9edd)
Adds ImagePullPolicy configuration to CSI node plugin and controller
plugin specs in both the operator config and driver CR generation functions.
This ensures the CSI driver components use the configured pull policy
from CSIParam.ImagePullPolicy.
Signed-off-by: Praveen M <m.praveen@ibm.com>
(cherry picked from commit 67c4203531)
ensure each gw pod runs on a different node when possible.
add default pod anti-affinity and topology spread constraints.
keep user placement overrides by merging defaults with spec placement.
add tests for default spread and user override behavior.
Signed-off-by: Oded Viner <oviner@redhat.com>
(cherry picked from commit bc6e5fbc86)