RDS MySQL instance runs our disk space temporarily

[Rajesh Kuppili]

Hi,

We have had two instances of our RDS MySQL instance run out of disk space for a few minutes, and then recovers back completely on its own. Any ideas what could be causing this?

Currently, I'm suspecting that this could be caused by temporary files created by a bad query, for which mysql might be creating a large temporary table, which is getting written onto disk. But I'm unable to confirm that, because when the RDS instance runs out of disk space, it doesn't log slow queries either. Is there a way to confirm this? Assuming this is the root issue, one approach is to create the tmp disk on a separate, but that is not possible to be done on an RDS instance, since you don't get shell access. Any other way to limit total temporary table size on disk?

This has brought down our website twice today, and could get out of control quickly. Need urgent help!

Thanks in advance.
Rajesh