[volunteers] ...and restarted Mailman's qrunner

[Rick Moen]

Last night, when I noted that Exim4 (SMTP server) wasn't running on
lists.svlug.org and started it, I was very short of time.  I _did_ 
check the process list[1] and made sure the other key processes were
running:  qrunner, spfd, spamd, httpd.  And I saw a brief flurry of
delayed postings hitting the main SVLUG mailing list, plus my posting to
this mailing list saying I had just started Exim.

Well, today, I happened to post to the main SVLUG list (Kindle irony
alert), and my post didn't show.  So, I checked the Exim logs:  The post
was duly received and passed along to Mailman qrunner -- which didn't
seem to be processing it.

So, I did _then_ what I probably should have done last night:  I shut
down Mailman, then checked ps to ensure that it was really purged from
running processes, then started it again.  And lo!  Postings started
going through again.

Moral:  If exim is wedged or bombed out of memory, and you need to
restart it, you really should bring down _all_ of the mail-handling
daemons -- qrunner, spfd, spamd, exim -- make sure they're actually
stopped[2], and then start them again, with exim being the last one you
bring up.  It's not safe to assume that the other mail-handling process
are OK, if you know one of them got borked.

[1] Again, you must do this as the root user.  Because of grsecurity,
the machine's "ps" command shows non-root users their own processes
only, and silently omits all others.

[2] That is, don't just trust to the SysVinit script to do shutdown: 
Do "ps auxw | grep exim" to make sure it really did get stopped.