Hello, it is about three days that a report is sent about exceeding the io rate.

I've read that I should use iotop, but the issues last for about one hour every day (at different timings) and I had never the chance to be there at the right time. So far.

What surprises me is that the CPU and network stats have standard values, and there are no cron scripts doing backups or similar, so I can't see a reason about these peaks.
Is there something I can look until I can finally use iotop?
Thanks.

Look at your I/O graph in the Dashboard. Is it a spike in regular I/O, a spike in swap I/O, or both?

You can also let iotop run as a cronjob and write the log somewhere, to figure out what's going on.

Googled this, so not a Linode, but a Slicehost link!
http://articles.slicehost.com/2010/11/1 ... o-and-swap

@hybinet, the swap I/O looks unrelated, there are some minor bumps but not in all cases.

The graph can be seen here: https://docs.google.com/open?id=0B1rEqCGSzGg7ZzVsclBXRHJMX0k

Thanks Nuvini, I'll give it a try.

Assuming you are running a web server... you can easily get some high IO disk rates if you're not doing any caching. The memory and CPU will look relatively normal but the disk IO will be screaming.

Do you have anything in your crontab? (crontab -l)