[Mailman-Users] Checking for Heartbeat

[Mark Sapiro]

Rex Goode wrote:
>
>I host with jumpline.com, which generally does a good job with me. The 
>Mailman installation is pretty good, but for some reason I can't track 
>down, Mailman keeps going belly-up. Jumpline customer support doesn't 
>know why. I have to stop it to turn off whatever qrunners are running 
>and restart it. I don't mind doing that now and then. It doesn't happen 
>that often.


I'm guessing you have a VPS or similar hosting arrangement where you
actually have access to the Maulman installation.

When Mailman 'dies' are one or more qrunners missing? If so, look in
Mailman's error and qrunner logs to try to find out why. If not, maybe
there is a lock issue of some sort. See the FAQ at
<http://wiki.list.org/x/A4E9>.


>Here's the problem. My users don't think to tell me when they haven't 
>had any mail from the list for awhile. I just go along thinking all is 
>well until someone tells me, "Hey, Rex. We haven't had any email from 
>the list for a week." I restart and old email mostly catches up.
>
>I have an administrative page for all of my web sites and the various 
>things they do. I am on that page several times each day. I would like 
>to generate some message when Mailman isn't working, something I would 
>see there. A lot of the command line scripts seem to work even if 
>Mailman is not running. What can I use to check to see if it is alive?


I suggest you look at the mmdsr script. I don't know what you have
available in your installation, but if you have a Mailman contrib/
directory, you'll find it there. If not, go to
<http://bazaar.launchpad.net/~mailman-coders/mailman/2.1/files/head%3A/contrib/>
where you can download mmdsr and README.mmdsr.

mmdsr is designed to be run by cron at 23:59 daily and mail you a
report summarizing Mailman's logs, queues, etc. This is very helpful
in spotting problems.

If mailmanctl and some qrunners are running when mailman is not
working, and you want to check right now, you can look at

find /path/to/qfiles -print -name \*.pck

to see if any queues are backed up, and you can do

ps -u mailman --no-headers | wc -l

which should return '9' in a standard installation (mailmanctl and 8
qrunners).

You could incorporate one of these or something similar in your
administrative page.

-- 
Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan