issue: when a block-mode pvc is created with encryption enabled
there is some space reserved for the encryption metadata.
Which doesn't allows users to write extact amount of data that
they have requested for.
solution: create pvc with extra space needed for the encryption
metadata.
The extra space is added during the CreateVolume and ExpandVolume
operations. And while returning the response remove the extra space
so the client/user gets the requested size reported.
Signed-off-by: Praveen M <m.praveen@ibm.com>
added test for uncommitted changes in deploy directory under go-test GH action.
Also, created a new make target named `make check-deploy-committed` that
can be used to verify the uncommitted changes.
Signed-off-by: Nikhil-Ladha <nikhilladha1999@gmail.com>
This commit fixes the issue where the `csiCreationTimeKey`
field was missing during the rebuilding of the
`VolumeGroupJournalConfig` struct in the `Connect()` method,
which led to the `csi.creationtime` key not being stored in
the omap.
Signed-off-by: Praveen M <m.praveen@ibm.com>
After cloning the RBD snapshot, an rbdVolume is returned for the
CSI.Snapshot object. In order to use the rbdSnapshot.ToCSI() function,
the rbdVolume needs to be converted (back) to an rbdSnaphot.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Mergify matches `body-raw` as regular expression, but that makes it
difficult (if not impossible) to match the `[` in a string.
Fixes: #4751
Signed-off-by: Niels de Vos <ndevos@ibm.com>
When Jenkins jobs are started by a comment, and if the PR contains
`[skip ci]` in the description, Jenkins does not run the requested job,
nor set a status for the job in the PR.
This causes Mergify to add the `ok-to-test` label again, instructing a
GitHub Action to add comments to start jobs in Jenkins. Once all
comments have been posted, the `ok-to-test` label is removed. Mergify
then notices that the jobs were not run, and adds the `ok-to-test` label
again.... Endlessly looping of adding the label, commenting and removing
the label as a result.
By reporting the brokenness of the PR description and marking the PR as
Draft, the looping is prevented.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
There has been some confusion about using different variables for the
InstanceID of the RBD-driver. By removing the global variable
CSIInstanceID, there should be no confusion anymore what variable to
use.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Fedora 40 builds seems to require disabling seccomp. When enabled,
installing Golang inside the container fails.
See-also: containers/podman#21012
Signed-off-by: Niels de Vos <ndevos@ibm.com>
GitHub had issues with Fedora 40 when it was released. Hopefully this is
not the case anymore.
Closes: #4585
Signed-off-by: Niels de Vos <ndevos@ibm.com>
The Squid container-image that is used as base for the Ceph-CSI
container-image uses CentOS Stream 9.
Closes: #4659
Signed-off-by: Niels de Vos <ndevos@ibm.com>
updated the group stringer method
to have pool and namespace for
proper debugging/logging and to
use it with CLI as agrument as well.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
GetVolumeByID already returning detailed
error message, the caller just need to return
it. No need to add duplicate details to error
message.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
in ModifyVolumeGroupMembership RPC call,
flatten the required images before adding it
to the group or else if the parent is not
mirror enabled adding a child to the group
will fail.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
updating HandleParentImageExistence function
to return more details error which includes
the pool/namespace/image name
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit adds support for flattenMode option
for volumegroup.
If the flattenMode is set to "force" in
volumegroupreplicationclass parameters,
cephcsi will add a task to flatten the image
if it has parent before adding it to the group.
This enable cephcsi to then mirror such images
after flattening them.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
updated GetVolumeByID to return more
descriptive error so that caller no
need to add more details in
the error message.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit adds locks on reclaimspace operations to
prevent multiple process executing rbd sparsify/fstrim
on same volume.
Signed-off-by: Praveen M <m.praveen@ibm.com>
With the ControllerGetVolumeGroup operation the caller can verify that a
VolumeGroup exists, and validate the volumes that are part of it.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
There was a discrepancy between the objectId
when creating the lock and when releasing the lock
this caused every lock to hang.
Signed-off-by: NymanRobin <robin.nyman@est.tech>
It seems to be possible that the UUID was found, but the name is not
set. Checking on UUID makes the CreateVolumeGroup operation more
idempotent.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
The ModifyVolumeGroupMembership operation can be used to change the
volumes that are part of a VolumeGroup. Only empty VolumeGroups can be
removed, this operation is required to make that possible.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
A RBD image can only be part of a single group. While an image is added
to a group, check if the image is already part of a group, and return an
error in case it is.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Add extra error checking to make sure trying to create an existing
volume group does not result in a failure. The same counts for deleting
a non-existing volume group, and adding/removing volumes to/from the
volume group.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
