A lot of rblsmtpd processes

Started by lunaticds, May 02, 2008, 03:27:39 PM

Previous topic - Next topic

lunaticds

Hi All,

Don't flame the n00b...
I have a FreeBSD 6.3 box I've setup a mail toaster on, but I've come across an issue every few days where the server goes from working fine to not.
Machine will appear to accept email, and will allow POP3 connections to check mail boxes, but it won't deliver mail to those boxes. Checking the active processes shows me that there will be quite a few processes of rblsmtpd running, and the only way to get mail delivering again appears to be via rebooting the machine.

Machine only handles around 2000-3000 messages a day (98% SPAM), so it's load is not huge.

I can't any log files that appear to have any useful information in them.

... Another issue is that every 5 minutes Cron spits out "Another toaster-watcher is running,  I refuse to!
at /usr/local/bin/toaster-watcher.pl line 37" in an email.

Can anyone shed some light on these problems, where I might need to look, and what info I might need to find thats relevant enough to assist in finding a solution?

Cheers,
David

lunaticds

All,

I've created a new server sitting on FreeBSD 7.0-p2

Essentially I started having the same problem of  rblsmtpd threads start appearing after a day or so, and mail no longer being delivered to mail boxes. I would see dozens of them without any other useful detail.

After doing some research I started reducing the RBL servers down to two - and found that the problem was still occuring, but I could see the -r rblservername.xxx - thinking it may have been one of those servers, I removed both. and set rbl_enable to 0 to disable it.
By this time I'm finding that the problem is occurring every 4 hours or so, which is obviously a major issue. 

Having made this change, I find:
vpopmail 19627  0.0  0.1  5400  2612  ??  I     7:12PM   0:00.04 qmail-smtpd /usr/local/vpopmail/bin/vchkpw /usr/bin/true
vpopmail 19650  0.0  0.1  5400  2628  ??  I     7:12PM   0:00.04 qmail-smtpd /usr/local/vpopmail/bin/vchkpw /usr/bin/true
vpopmail 19667  0.0  0.1  5400  2636  ??  I     7:12PM   0:00.04 qmail-smtpd /usr/local/vpopmail/bin/vchkpw /usr/bin/true
vpopmail 19822  0.0  0.1  3804  2108  ??  S     7:13PM   0:00.02 qmail-smtpd /usr/local/vpopmail/bin/vchkpw /usr/bin/true

listed a few dozen times, and yet again, unable to get mail delivered to any inboxes.

The log files aren't showing me anything obvious.
Other research has suggested checking that clamav and freshclam are looking in the same location for the database, which I've done. It's also updating without any errors.

I was hoping after having discovered that a non-working RBL server might cause that rblsmtpd problem that this would resolve it. Obviously not.

Advice anyone?

Cheers,
David


pbrezny

#2
on the Another toaster-watcher is running, bit seems like you might have manually run toaster-watcher.pl while another run wasn't finished or something.

I've had success just removing the pid file in /var/run/ and killing off any hung toaster-watcher processes.

Not sure if that' the "right" way to do it but it worked here.

also check to make sure you dont' have a simultaneous cron running (for example executing toaster-watcher.pl both from /etc/crontab and from su root crontab -e)
Peter Brezny

purplecat.net

lunaticds

Thanks for the response pbrezny - I was starting to think I was lurking here all alone.

I've managed to get past the "another toaster-watcher is running" error when I made a fresh build.  Thought the checking cron entries might have been handy for the main problem I have - no luck there :(

Thanks for the advice on that one.

Now I'm just stuck with
vpopmail 19627  0.0  0.1  5400  2612  ??  I     7:12PM   0:00.04 qmail-smtpd /usr/local/vpopmail/bin/vchkpw /usr/bin/true
running rampant every 2 hours now, and not being able to receive email again until I reboot. Not much fun.

Cheers,
David

matt

If you have processes hanging around that long, then you most likely have a DNS problem. Or a config problem, where you have stale/dead RBLs in your config, causing long DNS timeouts.

Run the cron job by hand with the -v option:
   toaster-watcher.pl -v

if there's any long pauses, pay attention to where they are and fix them.