Description of problem:
After a forced-failover event from site A to site B, if you delete the image on site A (while it's still primary) via "rbd rm <image-spec>", rbd-mirror will correctly detect that it should replicate the image from site B back to site A. Unfortunately, it skips the actual image sync step and just creates an empty image.
Restarting the rbd-mirror daemon on site A or requesting a manual resync will correct the issue.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. rbd --cluster siteB mirror image promote --force <image-spec>
2. rbd --cluster siteA rm <image-spec>
A blank (zeroed) image will be re-created on siteA
A properly synchronized image will be re-created on siteA
Associated master branch PR: https://github.com/ceph/ceph/pull/14945
Fix was included under rebase to 10.2.10
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.