Varnish randomly stopped?

Events happening in the community are now at Drupal community events on www.drupal.org.
j0k3z's picture

My website went down today and I couldnt figure out what the issue was. Eventually restarting varnish brought the site back.

Any idea why varnish might have randomly stopped running?
Were can I find logs to see what the issue was?
What is the best way to monitor a server so I know when things go down or if resources running low?

Comments

re: Varnish randomly stopped?

Greg Coit's picture

I've not seen an issue with varnish crashing. Unfortunately the varnish log is disabled by default due to the amount of space it can take. You can enable it by uncommenting the VARNISHLOG_ENABLED=1 variable in /etc/default/varnishlog (If on AWS, I would first symlink /var/log/varnish to an EBS volume to keep the logs form filling the root filesystem).

You might also look to see if there's anything in the /var/log/syslog file indicating what happened.

We hope to soon start addressing server monitoring within the mercury project. There are many options - likley we'll be going with munin (similar to cacti). Varnish has some nice built-in monitoring apps - especially varnishstat.

Hope this helps,

Greg

--
Greg Coit
Systems Administrator
http://www.chapterthree.com

What is the best way to

neyvaz's picture

What is the best way to monitor a server so I know when things go down or if resources running low?

The "industry standard" would be Nagios. It will give you an early email/sms warning on resource limits of your machine (load, memory, disk, ...) and also on service up/down states (apache, mysql, website in general). However, nagios takes a day to set up for the first time.

Munin would be useful to graph that same data over time (memory, disk, mysql, ...). So you can see what went wrong. But munin will not issue any warnings. Also it will not react to any problems. Munin, however, is really easy to set up within the hour.

Both feature a web interface.

In your case, a simple website monitoring service would have done it, though.
http://www.google.de/search?q=website+monitoring

This happened on our server also

sreyas's picture

Hi,

Same issue happened on our server also.. Load on the server was just fine, apache and all the services were running quite good and I was able to access the website using domain.com:8080, but not as domain.com

After restarting the varnish service site is working fine.. am monitoring the server for now.

Regards
Sreyas

Mercury

Group organizers

Group categories

Post Type

Group notifications

This group offers an RSS feed. Or subscribe to these personalized, sitewide feeds: