| Previous | Next |
| ERROR_CSV_VOLUME_NOT_LOCAL | ERROR_CLUSTER_RESOURCE_VETOED_MOVE_INCOMPATIBLE_NODES |
ERROR_CLUSTER_WATCHDOG_TERMINATING
Watchdog termination is a recovery action after lost progress
ERROR_CLUSTER_WATCHDOG_TERMINATING indicates that a failover-clustering watchdog has decided a monitored process or operation is no longer making acceptable progress and is terminating it. This is different from an ordinary application exit: the termination is intended to break a hang that could prevent resource recovery or cluster state transitions.
Capture dumps and cluster logs before repeated restarts overwrite the evidence. Identify the monitored process, pending resource operation, callback duration, and thread stacks. Resource DLL entry points must honor timeouts, report pending status correctly, and respond to cancellation. Increasing timeouts without finding the blocked I/O, deadlock, or unresponsive dependency can extend the outage rather than fix it.
Hang evidence to collect
- Process and resource operation selected by the watchdog.
- Thread dump and wait chains before termination.
- Pending Online, Offline, LooksAlive, IsAlive, or control callback.
- Storage, network, service, or synchronization dependency blocking progress.
References
- Microsoft: reporting pending resource operations
- Microsoft: resource DLL timeout and callback practices
- Microsoft: cluster logs and failure-event correlation
Looking for a different code? Search another status or error code.