Ensure the function `ensureCustomResourceDefinitionsExist` returns when
the duration specified in `retryCRDIntervalMax` flag is reached.
This works by passing a context to the backoff so that we can
timeout with context.WithTimeout. The new (correct) default duration for
the flag is 30s to give a bit more time when controller and CRD are
created at the same time.
Also limit the amount of resources returned by the list calls to 0 to
speed up the operations because we only care that the CRD exists.
- Introduce new flag for volume group snapshots and run worker if flag is enabled
- Introduce the main controller for group snapshots in snapshot-controller
- Needed to start the http server outside of pkg/metric
- We needed this because we want to add other endpoints to the server
Signed-off-by: Grant Griffiths <ggriffiths@purestorage.com>
This patch adds two new parameters `retryIntervalStart & retryIntervalMax`
which can be configured to adjust the ratelimiters of snapshotqueue and contentqueue
in the controller.
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
```release-note
`retry-interval-start` and `retry-interval-max` arguments are added to common-controller
which controls retry interval of failed volume snapshot creation and deletion.
These values set the ratelimiter for snapshot and content queues.
```
Signed-off-by: Humble Chirammal <hchiramm@redhat.com>
This commit also adds the compatibility with dep (Go dependency
management tool) via symlink creation (same method applied in
kubernetes-csi/csi-test)
TODO: fix hack/update-generated-code.sh
Fixes issue #239.
Signed-off-by: Alex Szakaly <alex.szakaly@gmail.com>
