I am successfully (and happily) using courier-imap-4.0.6 (I use the 
Gentoo GNU/Linux distribution).

I had trouble using a courier-imapd account on my computer after it 
underwent a reboot. Whenever the client tried accessing the trash 
folder, courier-imap would sit waiting for a lock to be broken.

When my computer loaded its daemons after the reboot, the dkim-filter 
milter claimed the PID of a long-dead courier-imapd process. The process 
had been accessing the .Trash/tmp folder of this account right as my 
computer died, leaving the stale lock.

Is it possible for a future version of courier-imapd to correct such a 
problem itself by recognizing that /proc/$PID/exe != /proc/self/exe? I 
know that such a thing should not be checked everytime a lock is found 
because locking must be efficient. But if a lock can not be claimed for 
half a minute or so, I think the process would be benefited by examining 
the validity of the lock. I did not see this feature discussed in the 
ChangeLog ( http://www.courier-mta.org/imap/changelog.html ) for a newer 
version of courier-imap.

Also, is this the place to report bugs?

Below is an edited shell session of me finding and removing the stale lock:

margbr...@ohnopublishing ~/.maildir/.Trash/tmp $ ls -lha
...
-rw-r--r-- 1 margbrink margbrink    8 Jul 21 21:00 
1248224418.M734712P16152.ohnopublishing.net
-rw-r--r-- 1 margbrink margbrink   23 Jul 18 21:19 courier.lock
   $ less courier.lock
7892:ohnopublishing.net
   $ ps -Aef |grep 7892
milter    7892  7891  0 Jul19 ?        00:00:03 /usr/sbin/dkim-filter -x 
/etc/mail/dkim-filter/dkim-filter.conf -P 
/var/run/dkim-filter/dkim-filter.pid
1014     16669 16594  0 21:18 pts/6    00:00:00 grep --colour=auto 7892
   $ rm courier.lock


-- 
binki


------------------------------------------------------------------------------
_______________________________________________
courier-users mailing list
[email protected]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users

Reply via email to