| Previous | Next |
| hrBadBackupDatabaseSize | hrFeatureNotAvailable |
hrTermInProgress
Where the operation crossed the contract
hrTermInProgress means the instance teardown boundary: new work reached ESE after termination had begun.
This is the legacy Directory Service backup/restore HRESULT form of JET_errTermInProgress (0xC80003E8).
The key comparison is: this is later than hrNotInitialized and different from a transient table lock. The first useful observation is to record which component called JetTerm or stopped the owning service and which session or cursor was still active. This evidence shows whether startup, normal operation, shutdown, or restore ordering is the actual boundary.
Reconstruct the engine state
- all still-live sessions and child handles at the transition; associate it with this result rather than with a later generic exception.
- the first request accepted after shutdown or before initialization completed; associate it with this result rather than with a later generic exception.
- the successful or failed JetInit/JetTerm sequence and the component that owns it.
The ESE objects in play
| Diagnostic layer | ESE instance state and call ordering |
|---|---|
| Typical API surface | JetCreateInstance, JetSetSystemParameter, JetInit, JetTerm, and restore entry points |
| Code-specific condition | the instance teardown boundary: new work reached ESE after termination had begun |
| First corrective direction | stop admitting requests, close dependent cursors and sessions in ownership order, then initialize a fresh instance |
Handles derived from an instance are not reusable after teardown. Restore workflows have stricter ordering than ordinary open/attach calls.
Repair without destroying evidence
- Freeze the failing request context and record
0xC80003E8, the Jet API name, and the current instance/session ownership. - Apply the targeted fix: stop admitting requests, close dependent cursors and sessions in ownership order, then initialize a fresh instance.
- Before retrying the operation, reconcile instance state, admission control, and child-handle cleanup.
Do not collapse these codes together
This is later than hrNotInitialized and different from a transient table lock. The following neighboring results belong to the same broad subsystem but mark different boundaries:
hrNotInitialized | an ESE call requiring a running instance was made before successful initialization |
|---|---|
hrAfterInitialization | restore processing was requested after the ESE instance had already initialized |
hrInvalidOperation | the engine state, object type, or call ordering makes the requested operation invalid |
Actions that can hide or worsen the problem
- Do not reuse instance-derived handles after the lifecycle transition.
- Do not start a second initializer to work around an ordering defect.
Technical references
Looking for a different code? Search another status or error code.