When recovery master down, dlm_do_local_recovery_cleanup() only remove the $RECOVERY lock owned by dead node, but do not clear the refmap bit. Which will make umount thread falling in dead loop migrating $RECOVERY to the dead node.
Unified · Split
| @@ -2367,6 +2367,8 @@ static void dlm_do_local_recovery_cleanup(struct dlm_ctxt *dlm, u8 dead_node) · 2366 unmodified lines | |||
| 2367 | break; | 2367 | break; |
| 2368 | } | 2368 | } |
| 2369 | } | 2369 | } |
| 2370 | dlm_lockres_clear_refmap_bit(dlm, res, | ||
| 2371 | dead_node); | ||
| 2370 | spin_unlock(&res->spinlock); | 2372 | spin_unlock(&res->spinlock); |
| 2371 | continue; | 2373 | continue; |
| 2372 | } | 2374 | } |