1

I have a VPS running 32 bit Ubuntu (Precise Penguin) on a Linode VPS, and something is making runaway disk usage. I've received multiple I/O warnings from Linode (not 'warning' in the sense of "Stop it!" but "FYI; this may be a problem") about IO usage on disk. I've pruned many of the larger things I had on the filesystem but didn't need, but that is just masking the problem. The disk keeps getting full, even if I delete my website logfiles (and my website is relatively high-traffic, but not nearly large enough to swamp a disk).

My present best attempt is to run (as root):

find / -print | xargs du -sh > ~jonathan/tmp/size.1; sleep 7200; find / -print | xargs du -sh > ~jonathan/tmp/size.2

and then, when that's done, run a diff. That process is running now.

Is there anything else I should be aware of? Are there known causes of runaway disk usage on Precise and/or Linode VPSes? Linode said they were not able to report what it was that was growing rapidly enough to trigger I/O warnings.

Thanks,

Shane Madden
  • 114,520
  • 13
  • 181
  • 251
Christos Hayward
  • 1,162
  • 3
  • 16
  • 35

2 Answers2

0

It could be anything you're running. Use iotop to see what process is doing the I/O.

womble
  • 96,255
  • 29
  • 175
  • 230
0

My Linode gave these sort of i/o usage warnings when I first set it up, whenever I was doing the migrations from the old server. I think it's the case of monitoring sensitivity than anything serious, if you're assured that your software is not misbehaving.

What I would do is look at the disk i/o graphs on the Linode dashboard, and set the monitoring levels to something that's appropriate for your usage. The default levels are too low.

cjc
  • 24,916
  • 3
  • 51
  • 70