Syncing latest changes from main for ramen #207
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit introduces changes to disown PVC resources when a VRG is deleted, particularly in scenarios where the DR is disabled. The motivation behind this adjustment is to give back the management of the PVC lifecycle to OCM. Upon disabling DR, the VRG now iterates through all volsync-managed PVCs. It removes VRG ownership from these PVCs and reinstates OCM annotations. This ensures that PVCs are not prematurely garbage collected upon VRG deletion. In order for that to work, a new annotation needs to be added to the VRG to indicate that the PVCs should be disowned upon DR disabling drplacementcontrol.ramendr.openshift.io/do-not-delete-pvc: "true" Signed-off-by: Benamar Mekhissi <[email protected]>
Signed-off-by: Benamar Mekhissi <[email protected]>
This commit changes the name and the values that we log with. The logger is created with the name of the controller. The logger also adds the values of customresource, its UID and the version. This makes it very easy to filter out the log messages based on those attributes incrementally, thereby allowing one to look at the reconcile logs at the the lowest granularity. The reconcile exit also logs the time spent for reconcile. Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Elena Gershkovich <[email protected]>
Signed-off-by: Elena Gershkovich <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
The verbosity was useful maybe early in the project, it doesn't help much anymore. This change also makes the default value good enough for the CI, therefore don't override the env in the CI. Signed-off-by: Raghavendra Talur <[email protected]>
This also lets us have a default KUBEBUILDER_ASSETS file and simplifies the Makefile. Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
this refactoring helps to set metrics even though one of the cluster is down, at this time,the metrics value is set as 0 and alerts are fired. Signed-off-by: rakeshgm <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: rakeshgm <[email protected]>
Current check for metro actions that use isMetroAction passed in the to and from clusters to check equivalence of region values. This would succeed when to and from clusters are the same, which can happen when the preferredCluster and the failoverCluster are the same (Deploy->Failover->Failover). As a result in a regional DR policy, such a state would enter processing or waiting for fencing to take effect before proceeding with the Failover. This commit switches to the policy based metro determination instead of the to and from clusters as before to resolve the issue. Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Signed-off-by: Elena Gershkovich <[email protected]>
…ionDestination resource post successful PV/PVC restore The issue arises during successful restoration of PV/PVC when the reconciliation proceeds to set up the ReplicationSource without saving the restore status. Subsequently, if a failure occurs or Ramen restarts before ReplicationSource setup completion, the next reconciliation fails due to the previously deleted ReplicationDestination. To address this, upon restore completion, reconciliation will exit, ensuring status update. In subsequent reconciliations, the restore is bypassed, and the ReplicationSource setup step will start. Signed-off-by: Benamar Mekhissi <[email protected]>
Signed-off-by: Benamar Mekhissi <[email protected]>
Signed-off-by: Benamar Mekhissi <[email protected]>
Signed-off-by: Benamar Mekhissi <[email protected]>
Signed-off-by: Elena Gershkovich <[email protected]>
When the request for DRPC to delete is received, we remove the finalizer and then delete the metrics. The object is then processed for deletion. The API server may not delete immediately and it may take some time for garbage collection to complete the request. Ramen may get the reconcile call again for which ramen may get the DRPC object which is the deletion state. At this point in time, we try to set the metrics with zero value and alerts are being fired. The fix here is that metrics will only be set if DRPC is not being deleted Signed-off-by: rakeshgm <[email protected]>
Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Signed-off-by: Shyamsundar Ranganathan <[email protected]>
This helps determine using an MCV later if the VRG on the managed cluster was created by a manifest work from the current DRPC resource instance. In cases of hub recovery this will mismatch as the DRPC is recreated on the hub and hence would have a different UID and helps to not orphan VRGs on the managed clusters due to missing hub manifest work for the same. The change to add the UID in the initial manifest work creation ensures other parts of the code that fetch the manifest and update it are not impacted. IOW, if a manifest work exists for a VRG it is created by the current instance of the DRPC as manifest work resources are not recovered on a hub recovery or in a declarative creation of the DRPC. Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Adoption ensures that there is a ManifestWork that manages the VRG resource on the ManagedCluster. In the case of hub recovery, we would find VRGs on the managed cluster without any ManifestWork for the same on the hub. This change ensures that such missing ManifestWork is created on the hub by the current DRPC resource. It does this by storing the DRPC resource UID as an annotation on the created VRG, and on any discrepency ensures that the current DRPC creates a manifest work for the VRG on the new hub. The case for upgrade is also handled, where a manifest work may be present but without the DRPC UID annotation on it. Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Test covers UID adoption by faking returns from MCV and also pre-creating an MW for one of the managed clusters. This covers most cases where the VRG needs to be adopted due to upgrade cases, or due to it being orphaned. Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Signed-off-by: Shyamsundar Ranganathan <[email protected]>
Explain why we need Go 1.20 and how to maintain multiple Go versions so ramen can be built and tested while using newer default Go version. Signed-off-by: Nir Soffer <[email protected]>
When comparing PVs, skip comparing unset "Spec.ClaimRef.kind". This
breaks validation when using KubeVirt VM, and actual resources in the
system do not match the backed up resources in the s3 store. It is
correct to ignore unset kind since this is an optional field[1].
Previously we failed with:
Failed to restore PVs: failed to restore ClusterData for VolRep
(failed to restore PVs and PVCs using profile list
([s3profile-perf8-ocs-storagecluster]): failed to restore all
[]v1.PersistentVolume. Total/Restored 1/0)
And then the VRG will not make any progress. Now we consider unset
"kind" as equal and continue the flow normally.
[1] https://github.com/kubernetes/api/blob/f3648a53522eb60ea75d70d36a50c799f7e4e23b/core/v1/types.go#L6381
Bug: https://bugzilla.redhat.com/2262455
Signed-off-by: Nir Soffer <[email protected]>
This current rook version fails to take latest ci changes. This commit updates the rook version. Signed-off-by: yati1998 <[email protected]>
Submariner renamed the `Synced` condition to `Ready`[1]. Good change but it broke our tests. [1] submariner-io/lighthouse#1211 Signed-off-by: Nir Soffer <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Signed-off-by: Raghavendra Talur <[email protected]>
Earlier, we were logging the maintenance error and also setting the requeue flag but the error was not being returned to the reconciler. Signed-off-by: Raghavendra Talur <[email protected]>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: df-build-team The full list of commands accepted by this bot can be found here.
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR containing the latest commits from main branch