Great Circle Associates Majordomo-Workers
(January 1997)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Lock file not being removed in digest
From: Dave Wolfe <dwolfe @ risc . sps . mot . com>
Date: Tue, 7 Jan 1997 13:42:45 -0600 (CST)
To: brozen @ webdreams . com (Brock Rozen)
Cc: majordomo-workers @ greatcircle . com
In-reply-to: <Pine.OSF.3.95.970107140242.28635M-100000@webdreams.com> from "Brock Rozen" at Jan 7, 97 02:04:14 pm
Reply-to: Dave Wolfe <david_wolfe @ risc . sps . mot . com>

[ Brock Rozen writes: ]
> 
> If the .archive directory does not exist and mkdigest is run, then digest
> will abort in the middle and the .LOCK file in the digest work directory
> won't be erased...
> 
> It is possible to put some additional checks in somewhere around the abort
> section?
> 
> At minimum, a note should be put in the FAQ that before restarting
> mkdigest/digest the lock file needs to be deleted.

If the lock file isn't valid, shlock will remove it. Granted, the
determination of whether or not a lock is valid is a little lame in that
it checks to see if the process ID in the lock file is that of a valid
process, which can be true if the pid has been reassigned, worse yet
if it's reused for a long-running process, so a note about persistent
"Can't lock...lock is valid" messages steering users to look for stale
lock files is entirely appropriate. But I don't think it's any of
abort's business to be cleaning up locks.

-- 
 Dave Wolfe


References:
Indexed By Date Previous: Re: Possible bug in access_check
From: Dave Wolfe <dwolfe@risc.sps.mot.com>
Next: Re: bug in majordomo.pl
From: Dave Wolfe <dwolfe@risc.sps.mot.com>
Indexed By Thread Previous: Lock file not being removed in digest
From: Brock Rozen <brozen@webdreams.com>
Next: bug in majordomo.pl
From: Angel Talamona <actf@rita.unr.edu.ar>

Google
 
Search Internet Search www.greatcircle.com