| Previous | Next |
| ERROR_VRF_CFG_ENABLED | ERROR_SHUTDOWN_IS_SCHEDULED |
ERROR_PARTITION_TERMINATING
The target partition is shutting down and no longer accepts the requested operation.
ERROR_PARTITION_TERMINATING has decimal value 1184 (0x4A0). It reports a lifecycle race: the caller retained or discovered a partition handle, identifier, or endpoint, but teardown had already started before the requested access completed.
Interpret “partition” through the returning API
Windows uses partition terminology in more than one low-level facility, including virtualization and isolated execution environments. The code itself does not identify the owner. Preserve the component, API, partition identifier, host process, and operation type rather than assuming a disk partition or treating the result as filesystem corruption.
Common race patterns
- a management service submits work after a stop or delete request
- a worker dequeues an operation while another thread tears down the partition
- a client retries through a stale handle after the hosting service restarts
- host shutdown cancels new work before all callers observe the state change
- inventory data lags behind the authoritative partition manager
Evidence to collect
- partition or compute-system ID and its last observed lifecycle state
- timestamps for create, start, stop, terminate, and failing access
- caller process, request correlation ID, and API or RPC operation
- host service events and any preceding crash, update, or administrator action
- whether the identifier was cached across reconnect or service restart
Correlating control-plane and worker logs usually reveals whether termination was expected or initiated by an unrelated failure.
Correct handling
Stop sending new operations to the terminating object and drain or cancel queued work. If the product contract allows recreation, wait until termination is complete, create a new partition, and obtain new handles rather than reusing the old ones. For destructive commands, treat “already terminating” as progress only after confirming that the same target and requested final state are involved.
Retry boundaries
A tight retry against the same object cannot reverse its lifecycle. Retry only at a higher orchestration level that can rediscover current state or create a replacement. Put a deadline on teardown observation; a partition stuck indefinitely in termination needs separate host diagnostics rather than endless client retries.
Developer recommendations
- make partition state transitions atomic from the scheduler’s perspective
- invalidate cached handles as soon as termination begins
- use generation numbers when identifiers can be reused
- make stop and delete operations idempotent
- test simultaneous create, access, stop, and host-service restart paths
Example
A compute-management client lists an isolated workload and immediately sends a configuration update. Between those calls, another administrator deletes the workload. The update returns 1184. The client refreshes authoritative state, removes the stale object from its queue, and reports that the target was terminated instead of retrying the old handle.
References
- Microsoft: System Error Codes (1000–1299)
- Microsoft: Windows containers overview
- Microsoft: Hyper-V technology overview
Looking for a different code? Search another status or error code.