|
Description
|
When svc.configd(1M) notices repository corruption after it has initialized
successfully (i.e. the initial consistency check has passed), svc.startd(1M)
notices only that svc.configd(1M) has died, and restarts it. If the failure
happens because svc.startd(1M) is causing svc.configd(1M) to notice a problem
in the repository, this leads to an infinite loop of "repository corrupt"
messages.
Instead, svc.startd(1M) should notice the CONFIGD_EXIT_DATABASE_BAD exit
code, and invoke sulogin(1M).
|
