procmail
[Top] [All Lists]

possible mishandling of diskfull

2005-04-03 03:49:14
OS: Linux, Debian Sarge
Libc: 2.3.2.ds1-20
Procmail: 3.22-11
Fetchmail: 6.2.5-12
Mutt: 1.5.6-20040907

This has happened to me a few times but this is the first time it's
happened to me at a time where I can actually report it. Basically, I
lost 6 hrs worth of email this morning due to (apparently) the fact
that /var/mail was full and procmail was not handling it right. The way
I acquire my email is via fetchmail. It uses pop to grab the mail out of
a remote mailbox and uses procmail directly to store it in the local
mailbox. The configuration for this remote loremoteuserion is as follows:

poll host, port 110, protocol pop3 via pop3.host with no dns:
        user remoteuser is localuser here with password *** and wants:
                mda "/usr/sbin/sensible-mda %F %T '' localhost"
        with options:
                fetchall

sensible-mda is just a passthrough. A common wrapper in debian that
fires off the MDA of the users choice. In this case it is procmail. Some
links that may be of use to the curious below:

http://packages.debian.org/unstable/mail/sensible-mda
http://modular.fas.harvard.edu/docs/debian-packages/sensible-mda/

Anyway, fetchmail is told to log to syslog and so some snippets from my
logs below:

Apr  3 06:48:54 theirongiant fetchmail[1052]: 2 messages for remoteuser at host 
(7025 octets).
Apr  3 06:48:54 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:1 of 2 (2030 octets)
Apr  3 06:48:55 theirongiant fetchmail[1052]: MDA returned nonzero status 75
Apr  3 06:48:55 theirongiant fetchmail[1052]:  not flushed
Apr  3 06:48:56 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:2 of 2 (4995 octets)
Apr  3 06:48:57 theirongiant fetchmail[1052]: MDA returned nonzero status 75
Apr  3 06:48:57 theirongiant fetchmail[1052]:  not flushed

That's when things are going well. Next hit of the pop server:

Apr  3 06:50:04 theirongiant fetchmail[1052]: 4 messages for remoteuser at host 
(14181 octets).
Apr  3 06:50:04 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:1 of 4 (2030 octets)
Apr  3 06:50:05 theirongiant fetchmail[1052]:  flushed
Apr  3 06:50:05 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:2 of 4 (4995 octets)
Apr  3 06:50:06 theirongiant fetchmail[1052]:  flushed
Apr  3 06:50:06 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:3 of 4 (3930 octets)
Apr  3 06:50:06 theirongiant fetchmail[1052]: MDA returned nonzero status 75
Apr  3 06:50:06 theirongiant fetchmail[1052]:  not flushed
Apr  3 06:50:06 theirongiant fetchmail[1052]: reading message 
remoteuser(_at_)pop3(_dot_)host:4 of 4 (3226 octets)
Apr  3 06:50:06 theirongiant fetchmail[1052]: MDA returned nonzero status 75
Apr  3 06:50:06 theirongiant fetchmail[1052]:  not flushed

In the meantime, I'm sleeping and nothing else is writing to the
partition. Mutt IS opened but it's not doing any writing. At most
it's checking for new messages and reading headers. Whilst doing so 
it puts a lock on the mailbox anyway so the two should be coexisting
nicely.

All in all it's kinda weird. I had thesame issue when fetchmail was
delivering locally to sendmail which was then invoking procmail as
above.

Any help would be appreciated. :)

-- 
    "To the extent that we overreact, we proffer the terrorists the
    greatest tribute."
        - High Court Judge Michael Kirby

____________________________________________________________
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

<Prev in Thread] Current Thread [Next in Thread>
  • possible mishandling of diskfull, CaT <=