Friday, November 6, 2009

Power failure - Restart not successful - No Email

There was a power failure here yesterday and knocked out the computers that were running.

One of these was my Internet server. It provides my web sites 'www.thegiblins.com' and 'shankillweather.com' and acts as my email server.

It is based on Linux, so, I thought that it should recover fairly easily.

This was not 100% true. The web site came up fine but email was working well.

I use sendmail as my MTA, which can be very difficult to configure and maintain.

I noticed that I hadn't received any emails since the afternoon. There are always emails being sent, either externally or internally from the various scheduled jobs that are always running. Something was wrong.


On further inspection of the /var/log/maillog, I say many messages like:

Milter: data, reject=451 4.3.2 Please try again later

and

to=, delay=00:00:04, pri=31034, stat=Please try again later

and even

to=root, ctladdr=root (0/0), delay=00:00:02, xdelay=00:00:02, mailer=relay, pri=30341, relay=[127.0.0.1] [127.0.0.1], dsn=4.0.0, stat=Deferred: 451 4.3.2 Please try again later

I said to myself, "What the Feck!".

I did the usual quick fix, i.e. "Did you try turning it off and on again?". No change.

I run Spamassassin and Clamav (Virus Protection), which can be troublesome boys, so, they were the next suspects.

The main clue from the maillog, was the word 'Milter'. This is basically a mail filter. There is one for Spamassassin and Clamav.

If you have ever set these up on Linux, you will know how tricky it can be. The documentation is not clear enough. It becomes clear after you have completed the job, which is not very helpful. Clamav has many components and it is not clear what are required.

Spamassassin seemed to be fine. It was identifying spam and non spam correctly.

Next was Clamav. I checked the configs, all OK. Ran an update on the data ans s/w. No change.

Checked my old notes. No help there. Talked to DR Google. Same.

I had a look through the bash history and noticed something.

I could see that 'clamav-milter' was running bit 'clamd' was not.

I started 'clamd' and restarted 'clamav-milter'. All good.

Done.

Mail logs looked good. Emails were coming through. Happy days.


What did I learn?

Make better notes (probably never happen).

Fully automate for a restart. I remember when I recently rebuilt the Internet server, there were issues with these email components. I obviously did not fully complete the install and left things not starting on boot.

No comments:

Post a Comment