Posted on 1 Comment

When hardware and monitoring is important: diskspace

There was just a thread on the Freenode #mysql IRC channel with someone wanting to switch off and delete their binary logs. Why? Because they were short on diskspace.
Mind you, this was a production system, so generally it’s rather a bad idea to disable binary logging there, unless you really don’t value your data – but in that case you might as well just close down your shop now 😉
This is not about MySQL reliability, but hardware can and will fail, and all kinds of other things can and will go wrong over time.

While I appreciate the jam this person was in today: choosing between not being online at all, and disabling the binlog for now…. it’s so much better to prevent this. I actually hear about database servers running out of disk space quite often so this is a common event!
It’s something to keep a close eye on, for instance using Nagios, and have an alert blip if space drops below a certain threshold. It can also be useful (for space management, safety and in certain cases also performance) to have tablespace, InnoDB and the binlogs each on separate volumes.

Posted on 1 Comment

1 thought on “When hardware and monitoring is important: diskspace

  1. If you are on Linux, using LVM and a file system that allows online resizing (e.g. JFS, ReiserFS, XFS) is also quite a useful in cases where the file system is about to run out of space.

    But of course you need to be able to provide the additional disk space without having to shut down the system 🙂

Comments are closed.