Description of the problem:

In the case of a 3 worker node OSD cluster with one OSD disk per node, if two osd disks are replaced rapidly and the odf recovery applied immediately after, noderecovery will intermittently not restore one of the two disks, and noderecovery will proceed to (and be stuck on) StorageClusterFitnessCheck. Deleting and re-running node recovery (at least in the case of the screen shot below) will allow the noderecovery to complete successfully.

How reproducible:

Intermittent

Version:

• ocp 4.18
• quay.io/jordigilh/odf-node-recovery-controller-catalog:v1.1.0-rc.6