The Network People Support Forums

This morning I wake up to being paged about SMTP down.
/var/log/mail/smtp/current shows:

Quote:

@40000000411b69a8191fa9dc rblsmtpd: 209.152.178.64 pid 3201: 451 temporary RBL lookup error @40000000411b69b70060179c rblsmtpd: 209.152.178.64 pid 3204: 451 temporary RBL lookup error @40000000411b69fc3af0fb4c rblsmtpd: 209.225.8.188 pid 3216: 451 temporary RBL lookup error

turned on toaster_debug in /etc/toaster-watcher.conf, manually ran toaster-watcher.pl; a few of these:

Quote:

testing postmaster.rfc-ignorant.org.... good, we have 12 NS servers, we can go on. querying 2.0.0.127.postmaster.rfc-ignorant.org...query failed for postmaster.rfc-ignorant.org: NXDOMAIN at /usr/local/lib/perl5/site_perl/5.8.5/Mail/Toaster/Qmail.pm line 1097

It even noted that it updated the /var/qmail/supervise/smtp/run file:

Quote:

InstallQmailServiceRun: updating /var/qmail/supervise/smtp/run...done. (diff): 25c25 < exec softlimit -m 35840000 tcpserver -S -H -R -c10 -x /usr/vpopmail/etc/tcp.smtp.cdb -u 89 -g 89 0 smtp rblsmtpd -c -r dnsbl.sorbs.net -r sbl.spamhaus.org -r list.dsbl.org -r cn.rbl.cluecentral.net -r kr.rbl.cluecentral.net -r relays.ordb.org -r xbl.spamhaus.org -r sbl-xbl.spamhaus.org -r bl.spamcop.net qmail-smtpd /usr/vpopmail/bin/vchkpw /usr/bin/true 2>&1 --- > exec softlimit -m 35840000 tcpserver -S -H -R -c10 -x /usr/vpopmail/etc/tcp.smtp.cdb -u 89 -g 89 0 smtp rblsmtpd -c -r dnsbl.sorbs.net -r sbl.spamhaus.org -r list.dsbl.org -r cn.rbl.cluecentral.net -r kr.rbl.cluecentral.net -r relays.ordb.org -r relays.visi.com -r xbl.spamhaus.org -r sbl-xbl.spamhaus.org -r bl.spamcop.net qmail-smtpd /usr/vpopmail/bin/vchkpw /usr/bin/true 2>&1

Cron confirms that toaster-watch.pl had been running every 5 minutes.  So what happened? Isn't worst-case scenerio supposed to be that if an RBL goes down; we give 5 minutes of temporary error, then toaster-watcher fixes it, and we ignore that RBL from now on. /var/log/mail/smtp shows that it spanned atleast 30 minutes.



Upgraded from 3.44 to 3.45
(Matt, shouldn't the toaster upgrade instructions tell you to pkg_delete the old matt* and toaster* bsdpan entries; then perl Makefile.PL, make install, THEN toaster_setup.pl -s mattbundle ??)

-- then the problem mysteriously dissapeared.

what could be the problem Matt?

(Matt: why does this latest version of the toaster seem to rotate the @ logfiles every 5 minutes, only keeping a max of a few hours of logs on the system --

Quote:

-rwxr--r--  1 qmaill  qnofiles   2275 Aug 12 09:15 @40000000411b6d611215fb64.s -rwxr--r--  1 qmaill  qnofiles   4280 Aug 12 09:25 @40000000411b6fba0cd27e5c.s -rwxr--r--  1 qmaill  qnofiles   3503 Aug 12 09:30 @40000000411b70e5093c99e4.s -rwxr--r--  1 qmaill  qnofiles   1086 Aug 12 09:35 @40000000411b72110b84055c.s root:tract|9:35:11am|/var/log/mail/send:

as you can see, they're not violating the maxsize_bytes variables (of ~1MB))

I run a series of tests on each RBL. However, it's quite possible that the RBLs fail in some way that I have not previously considered.

Matt