bottlenecking?

[Julian Field]

Quoting Tim Tyler <tyler at BELOIT.EDU>:
> Mailscanner experts,
>    We are currently running mailscanner 2.6 on AIX4.3.3
> systems.  Occassionally, we have noticed that mailscanner clogs up at
> peak
> periods and can't seem to process email for hours.  This means that local
> or remote mail is stuck in /var/spool/mqueue.in for hours before
> eventually
> making it to the normal sendmail mqueue (if at all).   If I kill the
> mailscanner process and restart it everything gets processed nearly
> immediately.
>
>   Isn't mailscanner_check running in crontab supposed to make sure that
> the
> process is running ok?  It would really be better if it simply killed the
> existing process every time and restarted it.  I suggest this because if
> it
> thinks the process is running ok when it isn't, then its useless.
> Anyone's
> thoughts about killing and restarting the process every 20 minutes or so?
>
>   Are there any other suggestions for configurations that can minimize
> bottlenecking?

Not sure why you are hitting the problem. However, what you could do is set
the "Restart Every" interval much shorter. Then MailScanner will re-exec
itself more frequently, which will probably work around whatever the
underlying problem is. "Restart Every = 1200" would make it restart every 20
minutes.

--
Jules