Because of the recent CRD changes made, the mon count had a minimum of
1, making the configuration of the external cluster impossible. The
operator would fail to add the finalizer:
```
2021-04-29 16:57:38.429451 E | ceph-cluster-controller: failed to reconcile. failed to add finalizer: failed to add finalizer "cephcluster.ceph.rook.io" on "test-external": CephCluster.ceph.rook.io "test-external" is invalid: spec.mon.count: Invalid value: 0: spec.mon.count in body should be greater than or equal to 1
```
We now fixed the CRD as well as adding the CR status.
Signed-off-by: Sébastien Han <seb@redhat.com>
The schema in v1.6 is more strict and we are seeing errors in the csv
upgrades where nullable values are not allowed for the pool settings.
We now allow the nullable settings again to fix the upgrade scenario.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Try to mitigate the following error from the CI:
```
+ sudo vgcreate test-rook-vg /dev/sdb
Can't open /dev/sdb exclusively. Mounted filesystem?
Error: Process completed with exit code 5.
```
Now we retry up to 3 times to create the volume group.
Closes: https://github.com/rook/rook/issues/7487
Signed-off-by: Sébastien Han <seb@redhat.com>
We now the generate the CRDs as well as the CSV on a single GOARCH only:
amd64. This helps us avoiding builds colliding with each others.
Closes: https://github.com/rook/rook/issues/7655
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit adds information about static pv and static pvc
creation and links a document with more information.
Signed-off-by: Rakshith R <rar@redhat.com>
The following command ended up being missinterpreted on linux,
generating a wrong file named ceph.clusterserviceversion.yaml'\''
with broken quotes.
Making it an array ensures the command and the args are parsed
correctly. Tested on both linux and macos.
Signed-off-by: Sébastien Han <seb@redhat.com>
when store.Status was not initialized it caused
crashes when assigning store.Status.Info["endpoint"]
to objContext.Endpoint.
This commits makes sure store.Status is not Nil
before setting objContext.Endpoint and
objContext.SecureEndpoint in the ObjectStoreUser
controller.
Closes: https://github.com/rook/rook/issues/6916
Signed-off-by: Ali Maredia <amaredia@redhat.com>
set fsPreparedForUpgrade to true for upgrading mds, to
make sure that max_mds is set back to active count
after all mds agent gets upgraded
Signed-off-by: shenjiatong <yshxxsjt715@gmail.com>
When DEBUG logging is enabled in the operator, there are a number of overwhelming messages in the log that are very overwhelming and don't seem useful, which makes debug mode difficult to use
Updated and Cutted down the messages that are not useful in debug mode
Closes: https://github.com/rook/rook/issues/7499
Signed-off-by: parth-gr <paarora@redhat.com>
RGW supports only v2 version of kv engine, current integration test is written for kv v1,
hence seperating existing test into two.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
Add `SecuritySpec` to specify kms details for rgw in `ObjectStoreSpec` than using existing
`SecuritySpec` used for `ClusterSpec`. The vault secret engines conflict with OSD and RGW
kms encrytpion
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
The ValidateConnectionDetails() contains `clusterSpec` assuming `SecuritySpec` only
part to `cephCluster` CRD. But we may need to validate same for `SecuritySpec` in
`cephObjectStore` CRD. Hence modifying the signature of the api.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
OBC operations will fail if RGW only enabled HTTPS. With this change, OBC operations
will use needed TLS certs to succeed the operation.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
Rook operator sends S3 to RGW server for bucket operations performed by healthchecker,
bucketprovisioner etc. Add api to fetch cacert if the TLS enabled for RGW.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
Allow metadata on VolumeClaimTemplates in StorageClassDeviceSets to be
persisted. The latest changes to controller-gen allows us to embed
objectmeta for embedded types.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
The ceph upgrade blocks if the ceph heatlh is HEALTH_ERR. If the user
wants to force the upgrade despite the health error, they will need
to set skipUpgradeChecks: true in the cephcluster CR. This message
will hopefully make that setting more discoverable.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The nodeAffinity for portable OSDs is needed to ensure the OSDs are
tied to the same zone or other failure domain where the OSD was
prepared and where it is expected to run. During upgrade the topology
affinity was not being detected since the osd prepare job is skipped
in that case. Now the topology will be detected from the running
OSD daemon pod's node.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
we have lots of duplicate codes for
github action. this commit refactor
them to a bash script which will minimize
duplicate codes.
Signed-off-by: subhamkrai <srai@redhat.com>
When upgrading Rook clusters created before v1.5,
`kubectl apply -f crds.yaml` fails due to the existence of
the usupported `spec.preserveUnknownFields`. Add text documenting how to
work around this issue in the Ceph upgrade docs.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
Sébastien Han
Travis Nielsen
Satoru Takeuchi
parth-gr
Rakshith R
shenjiatong
Ali Maredia
Blaine Gardner
Jiffin Tony Thottan
Blaine Gardner
subhamkrai
rohan47