Version 2.52-1 released
dpalmer at SKIDMORE.EDU
Thu Oct 11 19:44:53 IST 2001
Afraid we had to back off to 2.51. Mailscanner was dying periodically -
although I'm not certain that check_mailscanner wasn't precipitating the
problem. MailScanner's been running fine since we reverted back to the
prior release - still have 'check_mailscanner' cron.
Julian Field wrote:
> At 14:59 11/10/2001, you wrote:
> >However, twice within the past 24 hours, mailscanner has become hung.
> >It's still running, but nothing is passing out of mqueue.in. The last
> >time it looks like it restarted at midnight, but never processed
> >anything after that.
> >When I came in this morning there were over 7000 messages waiting in
> >mqueue.in and only 70 in mqueue.
> >I killed mailscanner and ran check_mailscanner to restart. Everything
> >is moving right along now, although it will probably take a few hours to
> >catch up. This is running on an E250 Solaris 8.
> >Plenty of disk space available all around. Non-default options in the
> >config : Deliver Method = queue rather than 'batch', Notify Senders=
> >no, Notify Postmaster = no,
> >Deliver in Background = yes.
> > Any ideas regarding why mailscanner seems to be getting hung up?
> >Anyone else having this problem?
> I have seen this problem myself this morning, though I don't yet know why.
> The last thing in my log was "Going to scan 1 message", but, as you also
> say, the mailscanner process was still alive, just not doing anything. I
> hope to have time to wade through the code myself tomorrow, I need to add
> more debugging info to it as I cannot yet see why it would just hang. As
> far as I know at the moment, there's no reason for any particular bit of
> the code to just not terminate. Restarting it with exactly the same message
> queue solves the problem, so it can't be anything caused by a particular
> message. What I'm wondering is whether there are bugs in the implementation
> of POSIX process group code that I now use (which I didn't before the
> latest version). This was the only way of avoiding a timeout-failing
> problem with one user of Solaris 2.7.
> If you see this problem a lot, back off to 2.51 (it's still on the web
> site, there just aren't any links to it) and see how you get on.
> Julian Field Teaching Systems Manager
> jkf at ecs.soton.ac.uk Dept. of Electronics & Computer Science
> Tel. 023 8059 2817 University of Southampton
> Southampton SO17 1BJ
Deanne Palmer dpalmer at skidmore.edu
Systems Administrator phone: (518) 580-5914
Skidmore College fax: (518) 580-5936
Saratoga Springs, NY
End of network mail
More information about the MailScanner