The `watch` verb is not required for the csi-snapshotter sidecar
to function, removing it from the deployment
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
Ceph upstream telemetry (which Rook users could optionally enable via
Ceph CLI or Dashboard) does not easily identify if a cluster is managed
by Rook. Identifying Rook clusters is possible only if users also enable
the rook mgr module, which is optional and not often used.
To get better upstream insights into Ceph clusters managed by Rook, set
a new 'rook-version' field in Ceph's generic k/v store. This field can
be used to determine if a cluster is managed by Rook by its presence.
It additionally provides information about the Rook version used to
manage the cluster, which is another piece of potentially useful
information.
Regarding Ceph cluster identification via telemetry:
Ceph telemetry is anonymous and uses a different ID from the Ceph
cluster ID. Identifying the Rook version in telemetry does not add
directly identifying information. Therefore, there should be little
risk that adding the Rook version field will decrease anonymity of Ceph
clusters via telemetry.
Signed-off-by: Blaine Gardner <blaine.gardner@redhat.com>
Now rgw can be pass by FQDN too,
to avoid problems where it is required
The format should be similar to
Ip(<IP>:<PORT> format), <FQDN>:<PORT> format
Signed-off-by: parth-gr <paarora@redhat.com>
The startup probe for the OSD has been too aggressive to kill the OSD
in case the OSD is taking some time to start. The OSD may be self-optimizing,
scrubbing, or some other internal operation before it is ready to start.
Rather than disable the startup probe completely, the default is now
to retry for two hours in case the OSD is performing those operations.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Currently it was mention we are updating the csi-user
But more precisely we go and update the keyrings
which auto update users so replacing it
Signed-off-by: parth-gr <paarora@redhat.com>
The ceph-external-cluster-details-exporter.py
generates the json output even with
rgw-endpoint IP from different cluster
Added a check validate_rgw_endpoint which will validate
the above rgw_endpoint Ip by matching the cluster fsid
Signed-off-by: parth-gr <paarora@redhat.com>
We don't need to run any `cp` command since a destination path already
exists. Using `--no-clobber` is not needed and it results in a noop
anyways.
Instead, let's print the reason why we exit the container.
Signed-off-by: Sébastien Han <seb@redhat.com>
This commit adds psp RoleBindings to rook-ceph-rgw role and
rook-ceph-purge-osd role. Without this commit, I cannot perform
rook-ceph-purge-osd job on PSP-enabled cluster.
Signed-off-by: Yuichiro Ueno <ueno@preferred.jp>
we should iterate older clusters spec instead of
new cluster spec since in case of new storageClassdeviceset
added in spec. The total count of the new clusters and older
spec cluster is will not be the same and we'll get panic.
Closes: https://github.com/rook/rook/issues/10222
Signed-off-by: subhamkrai <srai@redhat.com>
When detecting the device property the filesystem was not passed but
used later to validate if the device should be taken or not. Now we read
the filesystem info from "lsblk" and populate it in the device type.
Closes: https://github.com/rook/rook/issues/9470
Signed-off-by: Sébastien Han <seb@redhat.com>
- disk UUID should only be got for "disk" type device.
- It's not necessary to fail prepare job when failing to get disk UUID
- We should consider that `sgdisk` reports UUID even if there is no GPT.
Closes: https://github.com/rook/rook/issues/9948
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
Since Quincy is released, Ceph master builds will soon
be tagged with the reef v18 version. This adds the capability
for rook to check for the reef version of Ceph.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
If prometheus is already installed in a cluster, it may not be configured
to watch for third-party service monitors such as for Rook.
To help with this configuration, a link is added to the doc for
enabling the third-party service monitors.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
OSD on PVC is tested in many patterns but OSD on device is not.
It's preferable to add the following patterns for OSD on device.
- An OSD on a raw disk
- Two OSDs on a device
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
The following steps in CI scripts create corrupted GPT headers.
```console
$ sudo sgdisk --zap-all --clear --mbrtogpt -g -- "$DISK"
$ sudo dd if=/dev/zero of="$DISK" bs=1M count=10
```
It results in the failure of the succeeding `sgdisk --print "$DISK".
Here is an example.
```console
$ sudo parted /dev/sdb mklabel msdos
...
$ sudo sgdisk --zap-all --clear --mbrtogpt -g -- /dev/sdb
...
$ sudo dd if=/dev/zero of=/dev/sdb bs=1M count=10
...
$ sudo sgdisk --print /dev/sdb
Caution: invalid main GPT header, but valid backup; regenerating main header
from backup!
Warning: Invalid CRC on main header data; loaded backup partition table.
Warning! One or more CRCs don't match. You should repair the disk!
Main header: ERROR
Backup header: OK
Main partition table: OK
Backup partition table: OK
Invalid partition data!
$ echo $?
2
```
We can safely use the simple `sgdisk --zap-all "$DEVICE"` here.
It's not necessary to convert "mbr" to "gpt" because we'll make new GPT
labels just after this command.
Signed-off-by: Satoru Takeuchi <satoru.takeuchi@gmail.com>
Once we are done cleaning up the content of the encrypted osd, let's
close the main LUKS device.
This avoids having dm devices on the system after the cleanup.
Closes: https://github.com/rook/rook/issues/10181
Signed-off-by: Sébastien Han <seb@redhat.com>
The operator is very bursty and when starting up may need more memory
to reconcile all the CRs initially before it calms down.
The default of 256Mi was not sufficient for many clusters, so now
we bump up the limit to 512Mi.
This is a follow-up from the same change to the helm chart in
https://github.com/rook/rook/pull/10195.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
The operator is very bursty and when starting up may need more memory
to reconcile all the CRs initially before it calms down.
The default of 256Mi was not sufficient for many clusters, so now
we bump up the limit to 512Mi.
Signed-off-by: Travis Nielsen <tnielsen@redhat.com>
Travis Nielsen
Humble Chirammal
Sébastien Han
Blaine Gardner
Sean Johnson
Yuval Manor
parth-gr
Yuichiro Ueno
Madhu Rajanna
subhamkrai
Satoru Takeuchi
Alexander Trost