We now support the addition of the PVC that acts as a metadata device
for a given OSD.
For this, you need to create a new `volumeClaimTemplates`, its name must
be "metadata" otherwise, Rook won't pick it up.
A template will look like this:
```
volumeClaimTemplates:
- metadata:
name: data
spec:
resources:
requests:
storage: 10Gi
# IMPORTANT: Change the storage class depending on your environment (e.g. local-storage, gp2)
storageClassName: gp2
volumeMode: Block
accessModes:
- ReadWriteOnce
- metadata:
name: metadata
spec:
resources:
requests:
storage: 6Gi
# IMPORTANT: Change the storage class depending on your environment (e.g. local-storage, gp2)
storageClassName: gp2
volumeMode: Block
accessModes:
- ReadWriteOnce
```
We now map block and block.db directly inside the container instead of
running ceph-volume activate. This is much cleaner.
Closes: https://github.com/rook/rook/issues/3852
Signed-off-by: Sébastien Han <seb@redhat.com>
The owner references for the csi driver are causing the csi driver
to disappear when the garbage collector rebuilds its references.
To unblock production clusters we remove the owner references
until we can find the correct fix for the owner references.
Neither the apiVersion of v1 or core/v1 is working.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
This will likely happen when re-running an orchestration and
bootstrapping a new cluster with unclean drives.
The prepare pod won't prepare nor list existing disks, thus the code
will later fail on reading disks properties.
Closes: https://github.com/rook/rook/issues/4810
Signed-off-by: Sébastien Han <seb@redhat.com>
In the case where the node disappears abrubtly,
there was nothing to trigger the nodedrain reconciler to clean up
the stale canary.
Signed-off-by: Rohan CJ <rohantmp@gmail.com>
When the annotations are generated for a ceph daemon,
the all annotations are merged with the component annotations.
If there are no all annotations, the component annotation
calculation was crashing. Now we check for the case where
the all annotations are not defined.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The CSI driver was disappearing in some clusters after certain events such
as restarting the operator or master nodes being restarted. This is due to
an incorrect owner reference on the csi driver back to the configmap. The
owner reference must specify core/v1 instead of v1 for the APIVersion
of the configmap. This issue is only hit when the k8s garbage collector
rebuilds its cache, which would happen at indeterminate times from
Rook's perspective.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Currently there is no option to turn on and off the
log level verbosity in csi containers, This PR adds
the functionality to select logging levels for csi containers.
Fixes: #4690
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
If we detect a partition we should not use it if the ceph version is not
at least 14.2.8 since it has the necessary changes to support partitions
with ceph-volume.
Signed-off-by: Sébastien Han <seb@redhat.com>
TuneSlowDeviceClass wasn't part of the StorageClassDeviceSet types but
only in VolumeSource which is not exposed in the CR spec so setting the
option would never work.
The only issue is that this apply to all volumeClaimTemplates
where each can point to a particular storage class...
Signed-off-by: Sébastien Han <seb@redhat.com>
Fixed conditions getting resetted after the operator restart.
Did the changes which required to implement conditions on the rook ceph cluster
Conditions will eliminate the current status.State and incorporates a type which
provides much more description to the current status of the cluster.
Signed-off-by: Nizamudeen <nia@redhat.com>
If for any cause the operator processing a mon failover is restarted, the
mon canary pods created in order to get the right placement for this
kind of pods can remain in <pending> state forever.
The following labels have been added to the mon canary deployments:
"mon_type" = "canary"
"canary_pvc_name" = <PVC name used for the canary pod>
A separate <removeCanaryDeployments> has been implemented for
removing the dependence of the variable <canaryCleanup>.
This function uses the <canary> labels to obtain the <real>
and <existing> list of mon canary deployments in any moment.
This change, by itself, will avoid most of the problems with
pending canary pods, but as second safe guard measure a call to this
function has been added in the health check, to <force> a canary pod
deletion if the cluster of monitors is healthy and complete (this implies
that no canary pods should be present)
Resolves#3764
[test ceph]
Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
14.2.7 will only be a CVE fix release so 14.2.8 is the release that will
contain the fixes for ceph-volume partition.
Signed-off-by: Sébastien Han <seb@redhat.com>
external-attacher v2.x is not compatible with v1.x
This PR makes the require changes in csi templates
and upgrade documentation for easily upgrade
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
Versions of k8s prior to 1.14 do not support the nullable field in
CRD properties. As k8s 1.13 remains part of our support matrix and
as this field must remain nullable, we must remove CRD validation
of SecurePort. Range validation of this field has been moved into
the logic of rgw.validateStore.
Fixes: #4693
Signed-off-by: Elise Gafford <egafford@redhat.com>
Code for pod anti affinity mechanism on ceph gateways is a duplica of the one used by ceph mon.
This is a refacto to use same code on both components
Signed-off-by: n.fraison <n.fraison@criteo.com>
Sébastien Han
Travis Nielsen
crombus
Rohan CJ
Maksim Nabokikh
Madhu Rajanna
Nizamudeen
Dmitry Yusupov
Juan Miguel Olmo Martínez
Sergey Kaydalov
Elise Gafford
n.fraison