| Previous | Next |
| ERROR_VMCOMPUTE_OPERATION_PENDING | ERROR_VMCOMPUTE_INVALID_STATE |
ERROR_VMCOMPUTE_TOO_MANY_NOTIFICATIONS
Notification callbacks are registrations with a finite lifetime
ERROR_VMCOMPUTE_TOO_MANY_NOTIFICATIONS points to callback management around a compute system, not to excessive guest log messages. HcsSetComputeSystemCallback registers a client callback for lifecycle and other HCS events, and each registration has an associated callback handle.
Repeatedly registering on reconnect, retry or status refresh without closing the previous callback can exhaust the supported registration count. The compute system may still be healthy while the management process has leaked notification registrations.
Evidence and corrective action
- Count callback registrations per compute-system handle and identify code paths that register more than once.
- Close callback handles during normal teardown, failed initialization and reconnection paths.
- Ensure callbacks do not recursively create new registrations while processing an event.
- Use one dispatcher to fan events out internally when several application components need the same notifications.
- Distinguish registration exhaustion from a burst of events; reducing guest activity does not release leaked callback handles.
References
- Microsoft: HCS event callback sample
- Microsoft: HCS API callback types
- Microsoft: Host Compute System overview
Looking for a different code? Search another status or error code.