In WBEM V1, whenever a standalone (local or remote) instance, class, method, or property provider shut down, CIMOM was not able to use this provider again. Starting with WBEM build 619, this problem is fixed.
Whenever a request to a provider returns an error code with a facility other than FACILITY_ITF (all documented WBEM error codes have this facility), CIMOM assumes that a system problem is preventing communications with a provider, releases the provider, reconnects, and re-issues the request. If the request fails again, no further attempts are made and the request fails.
This behavior was already implemented in V1 for event consumers. This behavior is not supported for event providers: once an event provider is down, it will not be restarted again