Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Backport for 2.7.4 #1186

Merged
merged 6 commits into from
Nov 14, 2024
Merged

Backport for 2.7.4 #1186

merged 6 commits into from
Nov 14, 2024

Conversation

mnecas
Copy link
Member

@mnecas mnecas commented Nov 14, 2024

mrnold and others added 5 commits November 14, 2024 11:22
The importer pod can get stuck creating if there are multiple networks
and the wrong default network override annotation is configured. This
changes the annotation as requested in
[MTV-1645](https://issues.redhat.com/browse/MTV-1645).

I reproduced the bug without the change, and managed a successful
transfer with the change applied.

---------

Signed-off-by: Matthew Arnold <[email protected]>
When performing a warm migration of a vSphere VM, the current version of
Forklift creates the snapshot for each precopy, gets the change ID for the
previous snapshot, and then deletes the previous snapshot. Per the VMware
documentation (see MTV-1656), the correct flow is to record the change
IDs for the old snapshot, delete it, and then create the new snapshot.

This patch implements the correct flow by adding vSphere-specific steps to
the warm migration itinerary to ensure that old snapshots are cleaned
up and change IDs are recorded before creating new snapshots. Due to
the change in flow it's now necessary to store the change IDs somewhere
between itinerary phases, so a list of disk/changeid pairs has been added
to the Precopy struct. This change requires the CRs to be regenerated.

Correcting the warm migration flow appears, at least in some
cases, to address the filesystem corruption documented in
https://issues.redhat.com/browse/MTV-1656.

Signed-off-by: Sam Lucidi <[email protected]>
@mnecas mnecas requested a review from yaacov as a code owner November 14, 2024 18:58
@mnecas mnecas changed the base branch from main to release-2.7 November 14, 2024 18:58
Issue:
When monitring the migration process the forklift requests the PVC from
DV ClaimName. On top of that the forklift checks the Annotation "cdi.kubevirt.io/storage.import.importPodName".
But this annotation is missing in the main PVC but is present in the
prime pvc, which format is "prime-$UID_MAINPVC". This annotation points
to the created pod which imports the volume.

Fix:
Search the prime PVC instead of the main PVC, so the forklift is able to
check the importer pod status.

Signed-off-by: Martin Necas <[email protected]>
@mnecas mnecas merged commit 062918e into kubev2v:release-2.7 Nov 14, 2024
3 of 5 checks passed
@mnecas
Copy link
Member Author

mnecas commented Nov 14, 2024

I have by accident included MTV-1645 Change default network annotation I removed it from 2.7 branch

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants