Mon failover reliability is improved in two scenarios with these changes.
First, if the failover is aborted because of failing to start the new mon
and wait for it to join quorum, the new mon will be fully removed and
the failover rolled back. Previously, the operator would leave the new
mon pod running even if it never joined quorum. Now the timeout for the failover
is increased to wait longer for the new mon, but if it times out, there
are no leftovers from the cancelled failover.
Second, if the operator is failed in the middle of a failover
before the old mon is removed, the operator needs to remove an extra
mon. The operator was only selecting a random mon for removal instead
of honoring the stretch cluster topology or looking for mons
on the same node that could be removed to restore the correct
desired number of mons.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
One can now use the operator configmap for CSI_CLUSTER_NAME variable
for setting ClusterName on the RBD images and CephFS subvolume.
This will be useful in cases like for example, when two container orchestrator
clusters (Kubernetes/OCP) are using a single ceph cluster.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
One can now tweak the operator configmap for CSI_ENABLE_METADATA boolean to
enable/disable setting of PV/PVC/PVCNamespace &
snapshot-name/snapshot-namespace/snapshotcontent-name on the RBD images
and CephFS subvolume.
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
As cephcsi v3.7.0 is released updating
CephCSI to v3.7.0 in Rook and defaulting
the minimum version to v3.7.0
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
we need use `!=` for string comparision in bash instead of
`-ne`. Also, need to correct periodicity if condition to
make it work as expected.
Signed-off-by: subhamkrai <srai@redhat.com>
Codespell had its first release in a year yesterday and added some
dictionary rules that are now being flagged in our code. Only one
issue required changes to our code, so modify the codespell config to
allow the words we want to allow.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
we are already deleting v1beta1 cronJob while creating v1 cronJob,
so no need to delete v1beta while v1 cronJob deltion.
Signed-off-by: subhamkrai <srai@redhat.com>
User can define his desired endpoint list in Zone CR so that it will
overwrite the default service name for rgw.
Resolves#6432
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
Signed-off-by: Jiffin Tony Thottan <jthottan@redhat.com>
Implement the SSSD sidecar design for NFS.
Because NFS documents are getting long, create an NFS
`Storage-Configuration` section with separated topics to keep the `NFS
Overview` document sane.
Adjust the SSSD design to allow any VolumeSource, not just ConfigMaps.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
The tool gofmt in go 1.19 requires certain formatting
in the comments section for better rendering.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The cephcluster.ClusterName is no longer available in the k8s api
with the ObjectMeta type, so we use the cephcluster.Name
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
During certain maintenance tasks the admin will own running
operations on the ceph mons and osds, and the operator should
not interfere with those operations. If the operator sees
any mon in debug mode, every reconcile and mon health check
will be skipped. Thus, mons will not be updated while any
one of them is in maintenance. During OSD reconcile, individual
OSD deployment updates will only be skipped for OSDs that are
actively being debugged.
The debug mode for osd and mon deployments is signaled by
creating the ceph.rook.io/do-not-reconcile label.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
CephFS CSI driver dont have/advertise controller publish/unpublish
capabilities, thus dont need attacher sidecar for its operations.
The presence of external-attacher adds on overhead and issues wrt
attachment in various scenarios. One of them would be the lack of
performance on syncing volumeattachment from api server..etc.
More or less we don't have controller publish and unpublish capabilities,
so we should not make use of this sidecar and cause
unnecessary addon here thus other issues.
similar changes have been added to CSI
https://github.com/ceph/ceph-csi/pull/3149
Signed-off-by: yati1998 <ypadia@redhat.com>
This commit introduces new log level called csidecarLogLevel
for the sidecars which are not maintained by the csi-driver.
This will allow to have different configuration options
for the sidecars like snapshotter, resizer, attacher and
provisioner containers which will help to avoid the
spamming of logs to an extent.
Signed-off-by: yati1998 <ypadia@redhat.com>
A Liveness sidecar container that is deployed
with the CSI pods will be helpful to check
csi driver is properly responding or not.
and there is no liveness or readiness
kubernetes probe added for this one.
Providing an option to enable this
sidecar deployment and disabling it by default.
fixes#8783
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The RGW support server side encryption with help of s3 protocol, till
now the `sse:kms` was support in which keys will be provided by the user
and but it will be saved in external management service like vault. Now
the support for `sse:s3` is added so the entire encryption key
management is performed by RGW itsels.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
In object package code, external rgw server check is whether ceph cluster
is external instead of rgw. Correcting such scenarios in the code base.
Signed-off-by: Jiffin Tony Thottan <thottanjiffin@gmail.com>
this commits add new field `MaxLogSize` inside `LogCollectorSpec` of
cephCluster cr which will take max size of log after which we want to
rotate the log.
Signed-off-by: subhamkrai <srai@redhat.com>
This commit replaces `os.Setenv` with `t.Setenv` in tests. The
environment variable is automatically restored to its original value
when the test and all its subtests complete.
Reference: https://pkg.go.dev/testing#T.Setenv
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
csi pods uses leader election and they were using
default values but seems like default values of leader
election are not working for some specific clusters.
Signed-off-by: subhamkrai <srai@redhat.com>
When the mgr services are updated by the operator to set the
active mgr properly, skip services that do not currently
have a selector set.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Some context canceled were missed and not reported correctly. We had
multiple reconcilers running in //. Since they all run in goroutine then
will continue to run until they timeout (if timeout is available,
otherwise would run forever until the op is restarted).
Signed-off-by: Sébastien Han <seb@redhat.com>
With octopus coming to end of life, we remove support from
Rook for deploying Ceph Octopus and assume a min version of
Pacific v16. Any checks for octopus or earlier are removed
from the reconciles since they are obsolete.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
In new deployments we always want to disable the insecure global id.
If the mgr deployment is newly created, we know it's a new cluster
and the mons are in quorum, so we can go ahead and disable the
insecure global IDs.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
At present the CephFS CSI driver works with default mode ie,
`ReadWriteOncewithFsType`. However `File` type is more apt for
the CephFS CSI driver and this commit bring that change. The
similar change is also introduced in ceph csi driver here:
https://github.com/ceph/ceph-csi/pull/3204
considering the `File` mode has been GAd in 1.23 kubernetes version
and also we are lifting one of the problematic code path in this
area via Ceph CSI driver changes, it is good to move the
fsgrouppolicy to `File`.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
earlier we're deleting v1beta1 always before creating
v1 job. Now, we delete only when k8s version is
latest(which supports v1 job) and v1 job is not created
which means this is the first time and this is the only time we'll
delete v1beta1.
Signed-off-by: subhamkrai <srai@redhat.com>
Kubernetes 1.24 and newer use a different path for staging the volume.
That means the CSI-driver is requested to mount the volume at an other
location, compared to previous versions of Kubernetes.
The backward compatibility should be taken care by the CSI driver.
See-also: kubernetes/kubernetes#107065
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Travis Nielsen
Prasanna Kumar Kalever
Madhu Rajanna
subhamkrai
zhucan
Blaine Gardner
Jiffin Tony Thottan
Joseph Lee
Mudit Agarwal
yati1998
Eng Zer Jun
Sébastien Han
shalev.p
Josh Soref
Humble Chirammal