Orphaned, undelivered files in mqueue.in

[Remco Barendse]

Yes, I really have only 4.11-1 running.

Strangely enough, immediately after increasing the Max Children = setting
back to 5 the problem disappeared and I had not had a single SIGHUP since!

That is the only thing I changed and immediately solved the problem.

On Thu, 9 Jan 2003, Julian Field wrote:

> At 20:36 09/01/2003, you wrote:
> >The thought crossed my mind, and it may not even be unlikely. I think I
> >found the cause of the problem. The time stamp of the orphaned files is
> >about the same as these messages in the maillog:
> >Jan  9 21:20:15 raveon MailScanner[30971]: MailScanner child caught a
> >SIGHUP
> >Jan  9 21:20:15 raveon MailScanner[30946]: MailScanner child caught a
> >SIGHUP
> >Jan  9 21:20:25 raveon sendmail[31137]: alias database /etc/aliases
> >rebuilt by root
> >Jan  9 21:20:25 raveon sendmail[31137]: /etc/aliases: 66 aliases, longest
> >10 bytes, 658 bytes total
> >Jan  9 21:20:25 raveon sendmail[31146]: starting daemon (8.12.5): SMTP
> >Jan  9 21:20:25 raveon sendmail[31151]: starting daemon (8.12.5):
> >queueing at 00:15:00
> >Jan  9 21:20:26 raveon MailScanner[31162]: MailScanner
> >Jan  9 21:20:27 raveon MailScanner[31162]: MailScanner E-Mail Virus
> >Scanner version 4.11-1 starting...
> >Jan  9 21:20:28 raveon MailScanner[31162]: Using locktype = flock
> >
> >But.... no reference to receiving any e-mail. I think sendmail is in the
> >process of receiving the e-mail, which is not finished, hence no entry in
> >the maillog and then MailScanner kills and restarts sendmail??
> >
> >Or does the sighup mean that something has crashed? I seem to have an
> >awful lot of those :(
> >
> >Every 5 minutes!!!!
>
> There is nothing in the code that would cause a SIGHUP every 5 minutes,
> something external must be causing that. The only time the parent process
> sends a SIGHUP to the children is when it is terminated with a "kill" command.
>
> I carefully added code to 4.11 so that these orphaned files would not be
> left behind. Can you just double-check you really have 4.11 and not 4.10
> fragments anywhere? 4.11 has a "deletes pending" list which it will execute
> if the process is HUP-ed during a "delete message" operation. So this
> really shouldn't happen any more....
>
> >If the sender's connection isn't all that fast it is very well possible
> >that the mail transfer indeed gets killed before transmission is
> >completed.
> >
> >I never had this SIGHUP problem with 4.10?
> >
> >On Thu, 9 Jan 2003, Julian Field wrote:
> >
> > > It is possible that you merely got them for non-MailScanner related
> > > problems, such as an SMTP client getting cut off half way through sending a
> > > message. Sorry if that sounds like I'm passing the buck...
> > >
> > > At 17:17 09/01/2003, you wrote:
> > > >This was the first thing I have checked. I took different parts of the
> > > >numbers of these df files and grepped the maillog for it. Strangely enough
> > > >I cannot find them. Also when browsing through the maillog and looking
> > > >around the same date/time as these orphaned files in some cases there
> > > >isn't any message whatsoever of mail being delivered / received.
> > > >
> > > >I have about 10 orphaned files in the mqueue.in, none in the outdir and
> > > >they are all from yesterday and this morning. I know for sure that I
> > > >didn't restart MailScanner last night or this morning. It's hardly
> > > >possible that MailScanner would restart itself at exactly the same time as
> > > >these 10 e-mails in 1,5 day? (This is a very low volume home server).
> > > >
> > > >On Thu, 9 Jan 2003, Julian Field wrote:
> > > >
> > > > > At 16:08 09/01/2003, you wrote:
> > > > > >Ever since I upgraded to MailScanner 4.11-1 yesterday I have several
> > > > > >orphaned files that are piling up in mqueue.in
> > > > > >
> > > > > >The only files sitting there are the df files, without any other
> > files.
> > > > > >Also these messages have never been delivered to the intended
> > recipient.
> > > > > >
> > > > > >Any ideas??
> > > > > >
> > > > > >Can I still get these df files delivered or extract them to make them
> > > > > >readable?
> > > > >
> > > > > Check both your /var/spool/mqueue and your maillog to see if the
> > message
> > > > > ids have already been delivered (or at least placed in the outgoing
> > queue).
> > > > >
> > > > > I thought I had fixed this in 4.11, but obviously not well enough.
> > > > >
> > > > > Do the times on the files correspond to times when you have done a
> > > > > MailScanner "reload" or "restart"?
> > > > > --
> > > > > Julian Field
> > > > > www.MailScanner.info
> > > > > MailScanner thanks transtec Computers for their support
> > > > >
> > > > >
> > > >
> > > >
> > > >--
> > > >This message has been scanned for viruses and
> > > >dangerous content by MailScanner, and is
> > > >believed to be clean.
> > >
> > > --
> > > Julian Field
> > > www.MailScanner.info
> > > MailScanner thanks transtec Computers for their support
> > >
> > >
> >
> >
> >--
> >This message has been scanned for viruses and
> >dangerous content by MailScanner, and is
> >believed to be clean.
>
> --
> Julian Field
> www.MailScanner.info
> MailScanner thanks transtec Computers for their support
>
>


--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.