We need a shim for consistency group support, which we may eventually
port to SDK but have not yet. Otherwise, this is rather straightforward.
Change-Id: Ic880b7a64cde2148c266d549c4768c669ba3f501
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Depends-on: https://review.opendev.org/c/openstack/openstacksdk/+/943800
The command line operation could never work due to the incorrect call of
the openstacksdk API. This is updated to make it work and report errors
back to the user.
Closes-Bug: #2116969
Change-Id: I87cc410853c03b00dd1549d67cb1b9a8145bcfaa
Signed-off-by: Doug Goldstein <cardoe@cardoe.com>
This is no longer assigned to novaclient, meaning we can use it for SDK.
Change-Id: I43d9ede39d36cc29301f94fa462b9b9d9441807c
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Currently the volume backup restore command returns with error
even though the restore is initiated.
This patch corrects the response received from SDK and processes
it in a human readable form.
Change-Id: I7f020631fbb39ceef8740775fd82686d90a6c703
Closes-Bug: #2063335
Depends-On: https://review.opendev.org/c/openstack/openstacksdk/+/931755
When we show a volume backup by name, it calls the get_backup
method in SDK which is only used for getting a backup by ID.
This patch modifies the approach to call find_backup method
which first tries the find by ID and then find by name logic
eventually returning the backup details.
Story: 2011234
Task: 51127
Change-Id: I926d8de9810fcf2e5335bbe35aaab15e1e36a5cb
The name of the errors change and we need to move things around a
little, but it's otherwise a straight swap.
Change-Id: I0a19765ebeaa14c0534faa1542165b76ed2bf4e2
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Well, mostly. We still keep our own flake8 hooks and the hacking hooks
enabled. Everything else can be handled by ruff.
Doing this enables a couple of hacking checks that were previously
unaddressed. It also highlights a few cases that flake8 missed. Both are
addressed.
Change-Id: If81c7055e9ef692425da2789bae18a96d04b104f
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
The incremental volume backup stopped working after we moved from
cinderclient to SDK[1]. This happened because SDK accepts the
``is_incremental`` parameter[2] rather than the ``incremental`` parameter
which is actually passed in the API request and was previously a valid
parameter for cinderclient.
This patch fixes the issue by passing the ``is_incremental`` field instead
of ``incremental`` from the OSC side which adds the ``incremental`` parameter
in the API request.
Request body after the fix:
'{"backup": {"name": null, "description": null, "volume_id": "<vol-id>",
"force": false, "container": null, "incremental": true}}'
[1] https://review.opendev.org/c/openstack/python-openstackclient/+/889748
[2] 10e5e20fc0/openstack/block_storage/v2/backup.py (L126-L134)
Closes-Bug: #2070080
Change-Id: I89bd3d2751267ec39f4dbd664b7873ab87a9ac6c
The ``volume backup list`` command does not show the
created_at column for backups which doesn't really
help the end-user since a critical part of a backup
is knowing when it was taken, and fast if there is
a rush doing recovery by restoring a volume from a
backup.
Change-Id: I8d8a7f36c468c9faa2c2c47bfa9ba9e1ca5b9858
This patch acts as a base framework to add the parameters needed
for manage volume support.
This includes 2 changes:
1. Move get_parser and take_action code to common methods which
can be utilized by both v2 and v3
2. Make _check_size_arg as a static method and move it inside
CreateVolume class since it's not used by other classes.
[2] was initially thought to be a follow up change but since we
are implementing changes into the _check_size_arg method for v3,
it makes sense to just include it in CreateVolume class to avoid
adding a new additional method. Similar changes are done for v2
as well.
Change-Id: I9315e457ebd6c5ba4cc67452f92c9dc8c139ee3c
This change is entirely automated save for the update of some mocks from
'io.open' to '__builtins__.open').
We are keeping this change separate from addition of the actual hook so
that we can ignore the commit later.
Change-Id: I0a9d8736632084473b57b57b693322447d7be519
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Most of these skips were unnecessary. The few that did generate warnings
could be skipped.
We also set 'skip_install' since there's no reason to build the package
for linting purposes.
Change-Id: I9644e5c19720b9c41c60e0a5882b7cd7f6a71f7b
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
The '--retype-policy' option is used in the 'volume set' command to
specify the migration policy during the retype operation. The
'--retype-policy' option does not convey the correct meaning of its
usage. The migration policy determines whether we are going to perform
the migration in the retype operation or not and is not related to the
actual retype which just changes the volume type of the volume.
Change-Id: I2ea8fd3f5277bb3422ccae915d05e8ad44ff1912
An openstacksdk 'find_foo' proxy method will return None by default if a
resource is not found. You can change this behavior by setting
'ignore_missing=False'. We were doing this in most, but not all cases:
correct the issue.
In the event of calling 'image delete' with multiple images, it will no
longer fail on the first missing image and will instead attempt to
delete remaining images before failing.
Change-Id: I1e01d3c096dcaab731c28e496a182dd911229227
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Currently 'openstack volume list' calls nova to resolve server UUIDs
to server names. This is not required if:
1. no volume is attached to an instance
2. no volume exists in deployment
This patch fixes this by checking volume statuses and, if any volume has
status 'in-use', we will call nova to resolve server names.
Note that we don't check for 'reserved', 'attaching', 'detaching'
states since those are transition states and doesn't guarantee that
the volume is actually attached to the instance.
Change-Id: Ic4d89db69244d3fba44d4b69c79b3e7632ee3d53
Add some pagination helpers to configure pagination parameters for
various commands. Two pagination schemes are supported, based on what we
currently support across OSC commands: marker-based pagination and
offset-based pagination.
Change-Id: I551bb4c3ff0568c6df5244a1d0f0669497bee58f
Signed-off-by: Stephen Finucane <sfinucan@redhat.com>
When creating a volume, the scheduler hints can be supplied as strings.
The "same_host" and "different_host" hints can also be supplied as a
list if affinity/anti-affinity to multiple volumes is requested [0]
The previously-used `KeyValueAction` only supplies strings as values - the
last one if multiple --hint contain the same key. An alternative already
used in `CreateServer` would be `KeyValueAppendAction`, but only a subset of
the scheduler hints accept lists, so we cannot use that in general.
Therefore, we create `KeyValueHintAction`. It contains both a
`KeyValueAction` and a `KeyValueAppendAction` object and calls the
appropriate action based on the beginning of the given values as defined
in `APPEND_KEYS`.
[0] d96b705774/cinder/api/schemas/scheduler_hints.py (L31-L65)
Change-Id: Ida7f4662dc9fea24510758541fd4f2622b73bf31
Another quality of life improvements. The key for this one is weird and
the whole thing is a little more involved, hence why it's kept separate.
Change-Id: I75aa85f27905104dc84fffe823c01b4c90a6a822
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
Add aliases for the enabling multiattach, caching and replication via
extra spec properties. These are useful since the syntax for setting
each of them is a bit weird and frankly not very discoverable.
Change-Id: I08b51224c66a34a9dcfcffc95847e61d9aac0e7e
Signed-off-by: Stephen Finucane <stephenfin@redhat.com>
