We use several VPS – Virtual Private Servers – to host a number of sites. Yesterday one of them came crashing down. How could that be when we knew we had over 300 megabytes of free space two days before?! It was time to do some sleuthing.

First off, we deleted all bloated log files. Might sound stupid – logs contain critical information -  but you can’t work with gigantic log files when pressed with time. What you need is to get the server back up and running. By the time we created some space by removing some of the access_log files from various directories, we noticed the space available was back to normal levels, but decreased by 3 or 4 megabytes per 10 minutes. Not a good situation.

More forensic analysis was needed. Time to watch the system closely. We soon discovered one error_log file would grow incredibly fast when one specific site was loaded in the browser. We immediately downloaded the log, opened it while it was still manageable and there was the problem: one specific plugin was creating error upon errors. Some wordpress table was likely damaged or dropped during a Wordpress upgrade so the plugin would send multiple requests to the server and bloat the error_log.

No time to investigate further, this was not an essential plugin, delete, delete and remove!
If you server starts crashing after little or no increase in traffic, investigate methodically, you’ll find out sooner or later the cause. In this case, a simple PLUGIN brought down an entire server by plugging the plumbing …