mds: Define kill points involved in clustered migration and recovery
We need to define all the separate points at which a break in 1) clustered recovery and 2) migration leaves a different on-disk state for the freshly-recovering daemons. Insert config-option kill points for each for testing purposes.
A lot of this is already done; we have eg kill_migrate_at stuff.
- Description updated (diff)
- Category changed from 47 to 90
- Assignee set to Patrick Donnelly
- Related to Feature #17835: mds: enable killpoint tests for multi-mds recovery added
- Category changed from 90 to Correctness/Safety
- Target version set to v13.0.0
- Tags set to luminous
- Component(FS) qa-suite added
- Target version changed from v13.0.0 to v14.0.0
- Tags changed from luminous to mimic,luminous
- Target version changed from v14.0.0 to v15.0.0
Also available in: Atom