My site hosted on Linode is going down frequently. Help me diagnose?
I'm trying to crowdsource some troubleshooting here as the guy who ran the backend side of our site has departed.
I have a site hosted on a Linode that is going down regularly recently. Error page (error 502) indicates that it is the host.
When I look at Linode Dashboard my graphs look like this:
~~http://i.imgur.com/4CyFyaa.png
~~http://i.imgur.com/KWydemE.png
The areas on the CPU graph where utilization is depressed to ~2% are when the site is not reachable.
Rebooting the Linode brings the site back online.
I'm not sure if this is relevant but Storage is 100% allocated to disk images. LISH shows:
~~http://i.imgur.com/vr98Fkp.png
Thank you in advance to anyone who can help me. ~~
1 Reply
The problem in that case, was memory usage. The server worked fine for days, eating memory, until it run out of memory and the linux kernel killed whatever process it thought problematic, usually it was MySQL that was killed. A reboot fixed the problem temporarily.
Based on the above charts, I don't see high cpu utilization, so please check your memory usage (free -t -h) and see how that evolves over time (check once or twice per day). You should also check the /var/log/messages, because any such action by the kernel will be logged there along with the name of the process that was killed.
You should also check other random items, like forgotten CRON jobs and custom scripts.
You did not mention what kind of website software you run, so maybe that plays a role. Check your web server logs for possible error messages, if Apache is returning 502 error (bad gateway) then it is possible that you are doing something wrong (messing about with ngnix and apache or whatever proxy in between). A messed up proxy configuration will produce 502 bad gateway errors.