| Summary: | net-mail/mailman init script fails to restart service correctly | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Brian De Wolf <bldewolf> |
| Component: | [OLD] Server | Assignee: | Hanno Böck <hanno> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | net-mail+disabled, sandino |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
Affects obsolete mailman2 package. |
The current net-mail/mailman RC script does not restart mailman correctly. This is due to two issues: 1) The stop command doesn't immediately stop mailman. 2) The start command (on the mailman side) checks if mailman is currently running and exits if it is. This means that openRC's attempt at restarting (stop then start) causes mailman to exit and never come back because it thinks it's still running. The behavior of #2 is detailed in the mailmanctl -h: -s/--stale-lock-cleanup If mailmanctl finds an existing master lock, it will normally exit with an error message. With this option, mailmanctl will perform an extra level of checking. If a process matching the host/pid described in the lock file is running, mailmanctl will still exit, but if no matching process is found, mailmanctl will remove the apparently stale lock and make another attempt to claim the master lock. Also, the RC script contains a svc_restart() function that would work but doesn't appear to be used by anything. It's unclear to me whether an RC script can override the openRC restart behavior.