The expected ceph version for exporter is changed temporarily because
some regression was detected in Ceph version 17.2.6.
Thus, disabling ceph-exporter for now until all the regression are fixed.
Signed-off-by: avanthakkar <avanjohn@gmail.com>
(cherry picked from commit 3e46522423)
use ClusterID of the k8s cluster while doing nslookup of the
exported mon or osd service IP
Signed-off-by: sp98 <sapillai@redhat.com>
(cherry picked from commit 9e8e85d68b)
if port was used by other daemon，the dashboard will start failed and the status of the ceph will be err，then we update the port to a unused port，the operator can not deal with it
Signed-off-by: zc <zc@zcdeMacBook-Pro.local>
(cherry picked from commit caf7b8c753)
PR #3143 added check to retain 6790 port for mons. Now
we can remove this check in Rook 1.11.
Signed-off-by: sp98 <sapillai@redhat.com>
(cherry picked from commit 755135582f)
Remove the `ceph-conf-dir` volume and volume mount from exporter
as it's no more required to spin up exporter daemon
Signed-off-by: avanthakkar <avanjohn@gmail.com>
(cherry picked from commit 1a000673a4)
The defaults in CRDs cause an update and trigger a new
reconcile if the value is not set. For the key rotation
setting, the default was only being updated when the kms
feature was enabled. This update was causing the reconcile
to fail in the kms tests. With the default now in code,
the default can remain empty and no cr update will trigger
a new reconcile unexpectedly.
Signed-off-by: travisn <tnielsen@redhat.com>
(cherry picked from commit cd94ed7ddc)
When msgr2 is required, ensure the mon endpoints passed to
the csi configmap are on port 3300. The mons will be
listening on both 6789 and 3300. Existing volumes can
continue using port 6789 while new volumes will use
port 3300.
Signed-off-by: travisn <tnielsen@redhat.com>
(cherry picked from commit 0484a3973f)
This commit adds util function to create key rotation cron job
such as getKeyRotationContainer, getKeyRotationCronPodTemplateSpec
and makeKeyRotationCronJob.
Signed-off-by: Rakshith R <rar@redhat.com>
(cherry picked from commit ef38aea2a4)
This commit adds functionality to be able to rotate
key encryption key of encrypted PVC backed OSDs.
Necessary changes such as adding update functionality
to kms and rbac changes are made as well.
Signed-off-by: Rakshith R <rar@redhat.com>
(cherry picked from commit 71e011f731)
The exporter requires the mon config and keyring.
Use the CEPH_ARGS env var to specify the config
instead of expecting a ceph.conf to be generated
by another component on the host path.
Co-Authored-By: Travis Nielsen <tnielsen@redhat.com>
Signed-off-by: Avan Thakkar <athakkar@redhat.com>
(cherry picked from commit b6013d9b11)
Add container ports for OSD container only when multi cluster service is
enabled in the cluster spec.
Signed-off-by: sp98 <sapillai@redhat.com>
(cherry picked from commit 12bbc9b1a8)
ceph crash pod gets rook version label added unintentionally, which is only for deployment label.
changes include:
1, make a copy of the labels for pod spec from deployment labels just before adding rook version label.
2, add unit test to make sure it works as designed.
This fix along with pr 11674 is to stop ceph daemons pod restarting unnecessarily after upgrading rook operator.
Signed-off-by: Ben Gao <bengao168@msn.com>
(cherry picked from commit ed6ed25257)
If a kernel device name change happens and a block device file
in the OSD directory becomes dangling link, this OSD fails
to start continuously. This problem can be resolved by confirming
the validity of the device file and recreating it if necessary.
The original problem happened in host-based clusters. A similar
problem exists in PVC-based clusters. In this case, if a PV,
corresponds to an existing OSD, and points to a missing block
device file, the OSD pod fails to consume this PV. Although this
behavior is undesirable, I don't think this problem should be
handled in Rook. If doing so, we must re-create the existing PV.
Rook shouldn't do such work.
Closes: https://github.com/rook/rook/issues/10860
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
(cherry picked from commit 3a44a394b5)
The osd needs to bind to an ip that will allow both internal
and external access for rbd mirroring scenarios.
THe operator will create a service for each osd and
instruction the osd to bind to the pod ip while
publishing the clusterIP of the service as the
public IP.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 14e7225e0a18aff85ede1d476fcaddbbfbcbb732)
(cherry picked from commit d655c86b2f)
With the new mgr HA implementation (based on readiness probe) we
don't need the mgr sidecar (live-watch) anymore. This commit is
intended to remove all the related code.
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit ff75ec96b7)
After an update, the mgr pod will be restarted. Since the
readiness probe will indicate the standby mgr pod is never
ready, we need to check simply for the running mgr pod
and not whether or not it is ready.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
(cherry picked from commit 4d1017f805)
few functions got change as they were deprecated
for ex: ioutil.Readfile change to os.Readfile
ioutil.TempFile change to os.CreateTemp
And fixed golang-ci-lint-issues
Signed-off-by: parth-gr <paarora@redhat.com>
(cherry picked from commit a84daf9bf0)
This fixes the the MultiClusterDeploySuite CI failure. The operator
is timing out waiting for the mgr deployments to be
ready, according to the WaitForDeploymentToStart() method. After
the reconcile times out after about five minutes, the next
reconcile succeeds since the wait is only done for new mgr
deployments.
Closes: https://github.com/rook/rook/issues/11685
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
(cherry picked from commit b464428978)
This commit adds code to supply variables required
to enable read affinity for rbd volumes.
It leverages `read_from_replica=localize,crush_locations=key:value`
krbd map option from ceph to serve reads from the most local OSD.
refer:
https://docs.ceph.com/en/latest/man/8/rbd/#kernel-rbd-krbd-options
Signed-off-by: Rakshith R <rar@redhat.com>
(cherry picked from commit 8b9ded1f77)
Set the rbd map default option to ms_mode=prefer_crc
when v2 ports are used and encryption is disabled.
If we dont pass/set the ms_mode in the map options
the kernel client will assume its using v1 ports
and map operation will fail.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The idea behind this change is to use the readiness probe to implement
the mgr HA mechanism. In the current ceph mgr implementation only the
active instance offers the command 'mgr_status' through the admin
socket. We use this command combined with a Readiness Exec Probe to
detect which manager is active. Kubernetes will automatically mark
it as ready and redirect any service traffic to the active instance.
Closes: https://github.com/rook/rook/issues/11640
Closes: https://github.com/rook/rook/issues/11638
Signed-off-by: Redouane Kachach <rkachach@redhat.com>
The rook-version tag must not be set on the pod spec labels
since it will result in the ceph daemons being restarted every
time there is a rook version update, even if the ceph version
or pod spec was not otherwise updated.
The rook-version tag was being added to the OSD pod spec
due to a shared pointer to the labels. The code intended
to only add the version label to the deployment labels,
but since the pod labels shared the map variable, the
pod unintentionally also had the version label added.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Mons that are configured on msgr1 will be using the v1 port
in the csi driver since only one port is allowed in the configmap.
To allow mons to be started on v2 while keeping quorum with the
v1 mons, the v1 mons must continue running on the v1 port.
So even if msgr2 is required, allow existing v1 mons to continue
running on that port.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The default configmap does not have any labels or annotations
applied. If the helm chart is later updated to configure the
ceph config overrides, the configmap will fail to be created
by the chart since it doesn't have the needed label and helm
annotations so helm will know it can take over the resource
management.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
avanthakkar
sp98
zc
xiaobaowen
Travis Nielsen
Thomas Way
parth-gr
Rakshith R
Redouane Kachach
Ben Gao
Satoru Takeuchi
Madhu Rajanna
Avan Thakkar