It's a hardmount. NFSv3 over TCP.
If I create the lockfile manually, the delivery stops to the mailbox, so
I'm sure I don't have a rogue server either.
All servers show the "can't delete lockfile", so it's not related to a
single one.
I can't tell if it happens only between servers only, or if I have the
issue also between procmails on a single box.
I'll try rewriting the code to insert unique ID into the lockfile and
check for it if nobody comes up with a solution :(
P.
On Fri, 8 Dec 2006, Bart Schaefer wrote:
Date: Fri, 08 Dec 2006 07:46:32 -0800
From: Bart Schaefer <barton(_dot_)schaefer(_at_)gmail(_dot_)com>
To: procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
Subject: Re: dotlocking issues over NFS
On 12/7/06, Paul Sandys <myj(_at_)nyct(_dot_)net> wrote:
I disabled attribute cache (acregmin=0,acregmax=0) on all nfs mounts, but
it made no difference.
I presume since you said that only dotlocking is enabled that these
are "nolock" NFS mounts? Are they also "hard" mounts? Even
procmail's careful lock file creation algorithm is not guaranteed to
work on a "soft" mount.
If it's a hard mount with attribute caching disabled, procmail's
algorithm should be reliable.
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail
Paul Sandys
network operations manager
http://www.nyct.net/
212.293.2620
____________________________________________________________
procmail mailing list Procmail homepage: http://www.procmail.org/
procmail(_at_)lists(_dot_)RWTH-Aachen(_dot_)DE
http://MailMan.RWTH-Aachen.DE/mailman/listinfo/procmail