Description of problem: There is a possible infinite loop within journal:tag_list class method. If a peer cluster is offline and not replicating journal entries from the primary image and the primary image's exclusive lock is acquired at least 64 times during this outage, when the rbd-mirror in the peer cluster attempts to replay the image, it will cause an infinite loop in the primary cluster's OSD. Version-Release number of selected component (if applicable): 12.2.x How reproducible: 100% Steps to Reproduce: 1. Mirror an image to a secondary cluster 2. Stop the rbd-mirror daemon 3. Run 'rbd bench-write' for a few bytes more than 64 times 4. Restart rbd-mirror daemon Actual results: The 'rbd bench-write' or 'rbd-mirror' daemon will cause the OSD to enter an infinite loop. Expected results: The OSD does not enter an infinite loop. Additional info:
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. https://access.redhat.com/errata/RHBA-2017:3387