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>
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>
A VolumeGroup CSI-Addons object contains a list of CSI Volumes. A
ToCSI() function makes creating such a list much simpler.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Register the volumegroup controller as part
of rbd controller server to serve the volume
group RPC spec.
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
The rbd_types package was initially created with references to the rbd
package. And the rbd package references the rbd_types package. Having
rbd/types was not possible due to recursive imports. After cleaning up
the rbd_types package, it can be renamed to rbd/types.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
The Volume interface will make it easier to work with the rbdImage
struct, as the functions are cleaner defined. This benefits work that is
needed for VolumeGroups and other CSI-Addons procedures.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
In the future we'll introduce a more standard interface for objects like
Volumes and Snapshots. It is useful to have the context passed as 1st
argument to all functions of those objects, including their Destroy()
function.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
This commit adds support for flattenMode option
for replication.
If the flattenMode is set to "force" in
volumereplicationclass parameters, cephcsi will
add a task to flatten the image if it has parent.
This enable cephcsi to then mirror such images after
flattening them.
The error message when the image's parent is
in trash or unmirrored is improved as well.
Signed-off-by: Rakshith R <rar@redhat.com>
ensure a clean and isolated environment for testing purposes.
Signed-off-by: Mayank Pal <mayankpal9654@gmail.com>
ci: Use temporary directory for unit tests
remove err = os.Mkdir('/etc/ceph-csi-config', 0o600)
Signed-off-by: Mayank Pal <mayankpal9654@gmail.com>
ci: Use temporary directory for unit tests
remove err = os.Mkdir('/etc/ceph-csi-config', 0o600)
Signed-off-by: Mayank Pal <mayankpal9654@gmail.com>
ci: Use temporary directory for unit tests
remove if err
Signed-off-by: Mayank Pal <mayankpal9654@gmail.com>
golangci-lint reports these:
The copy of the 'for' variable "kmsID" can be deleted (Go 1.22+)
(copyloopvar)
Signed-off-by: Niels de Vos <ndevos@ibm.com>
This commit modifies a test case to check creation of
PVC-PVC clone of a restored PVC when parent snapshot
is deleted.
Signed-off-by: Rakshith R <rar@redhat.com>
This commit adds ParentInTrash parameter in rbdImage struct
and makes use of it in getParent() function in order to avoid
error in case the parent is present but in trash.
Signed-off-by: Rakshith R <rar@redhat.com>
Currently we are assuming that only one
rbd mirror daemon running on the ceph cluster
but that is not true for many cases and it
can be more that one, this PR make this as a
configurable parameter.
fixes: #4312
Signed-off-by: Madhu Rajanna <madhupr007@gmail.com>
This commit replaces the user implemented function
`CheckSliceContains()` with `slices.Contains()`
function introduced in Go 1.21.
Signed-off-by: Praveen M <m.praveen@ibm.com>
This commit removes the `topologyConstrainedPools` parameter
from PV volumeAttributes as it is not required.
Signed-off-by: Praveen M <m.praveen@ibm.com>
Everytime a connection is copied with the .Copy() function, it needs to
be destroyed once the object is not needed anymore. This was not done
consistently, a few more locations require the freeing of the connection
resources.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
By returning a connected rbdVolume in parseVolCreateRequest(), the
CreateVolume() function can be simplified a little. There is no need to
call the additional Connect() and detect failures with it.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Not all snapshot objects are free'd correctly after they were allocated.
It is possible that some connections to the Ceph cluster were never
closed. This does not need to be a noticeable problem, as connections
are re-used where possible, but it isn't clean either.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
Just like GenVolFromVolID() the genSnapFromSnapID() function can return
a snapshot. There is no need to allocated an empty snapshot and pass
that to the genSnapFromSnapID() function.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
if rbd storage class is created with topologyconstraintspools
replicated pool was still mandatory, making the pool optional if the
topologyconstraintspools is requested
Closes: https://github.com/ceph/ceph-csi/issues/4380
Signed-off-by: parth-gr <partharora1010@gmail.com>
The only encoding version that exists is `1`. There is no need to have
multiple constants for that version across different packages. Because
there is only one version, `GenerateVolID()` does not really require it,
and it can use a default version.
If there is a need in the future to support an other encoding version,
this can be revisited with a cleaner solution.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
The `rbdGetDeviceList()` function uses two very similar types for
converting krbd and NBD device information from JSON. There is no need
to use this distinction, and callers of `rbdGetDeviceList()` should not
need to care about it either.
By introducing a `deviceInfo` interface with Get-functions, the
`rbdGetDeviceList()` function becomes a little simpler, with a clearly
defined API for the returned list.
Signed-off-by: Niels de Vos <ndevos@ibm.com>
This commit makes use of crush location labels from node
labels to supply `crush_location` and `read_from_replica=localize`
options during mount. Using these options, cephfs
will be able to redirect reads to the closest OSD,
improving performance.
Signed-off-by: Praveen M <m.praveen@ibm.com>
Implemented the capability to include read affinity options
for individual clusters within the ceph-csi-config ConfigMap.
This allows users to configure the crush location for each
cluster separately. The read affinity options specified in
the ConfigMap will supersede those provided via command line arguments.
Signed-off-by: Praveen M <m.praveen@ibm.com>
