fiasco, the colocated server that handles all email for mulix.org, appears to be having some qmail problems. I suspect Shachar’s anti-virus installation, as qmail complains about ‘temporary problems after the DATA stage’, which is exactly the point I’d expect the anti-virus to kick in.
I took a look at the server and restarted clamav and qmail, but that didn’t work. We have a pretty complicated setup and I’m afraid to mess it up worse than it is right now, so I guess it’ll wait until Shachar gets back 😦
[edit: should’ve thought of it earlier… you can always reach me at muli(at)il.ibm.com]
Muli Ben-Yehuda's journal 
oh dear… i just sent u an email a couple of hours ago. do you have an alternative email?
Comment by ideawerkz — April 17, 2004 @ 5:02 AM |
muli@il.ibm.com
I should get the original email as well, but it could take a while.
Comment by mulix — April 17, 2004 @ 5:38 AM |
Ah, good ol’ Qmail. I had to perform a lot of black magic (read: patching the source with random unofficial patches) this past weekend just to get it to actually reject invalid incoming e-mail addresses at SMTP-time instead of generating bounces (which does nothing useful except fill up the remote queue).
Comment by compwiz — April 17, 2004 @ 8:21 AM |
Just so people reading in retrospect know
It turned out to be a bug in amavis-ng. They used a perl SUID script with “#!/usr/bin/suidperl” instead of “#!/usr/bin/perl”.
I got the angry fiasco mob to agree to put the anti-virus back in, and to let me use fiasco as outbound relay (yippee! :-D) pretending it’s “for making sure we know as soon as a problem happens again”.
Email was down for less than 24 hours. Will be a lot less next time.
Shachar
P.S.
Linuxjurnal doesn’t have my name in the spell checker, but does have “yippee”…….
Comment by Anonymous — May 1, 2004 @ 9:15 AM |