1

I suspect that with my hard drives is something wrong.

iostat -d -x -m

root@srv2:~# iostat -d -x -m
Linux 2.6.32-5-amd64 (srv2.host.com)        04/23/13        _x86_64_        (2 CPU)

Device:         rrqm/s   wrqm/s     r/s     w/s    rMB/s    wMB/s avgrq-sz avgqu-sz   await  svctm  %util
sda               2.68   331.06    2.78   86.00     0.28     1.62    43.84     1.73   19.52   5.06  44.88

ATOP output (red colored):

DSK | sda | busy 78% | read 4 | write 118 | avio 64 ms |

Can anyone tell me, what is going on and what additional tests I need to do.

Improve this question

1 Answer 1

Reset to default
0

There are several thing you can do

  1. First dubble check that you have a backup if you think the hard drive is bad.
  2. Run iostat with parameters "iostat -d -x -m 1". Look for abnormal fluctuations. Look at top in another terminal and try see if the peaks in "await" and "svctm" is related to any process.
  3. Run iotop

  4. Run an iotest, e.g. iozone. Make sure you run synchronous io. If not, you might just measure the caches.

    iozone -R -i 0 -i 1 -i 2 -i 3 -i 4 -i 5 -i 8 -l 1 -u 10 -s 10m -r 8 -T -I

    has provided relevant test for me. If you have a magnetic disk without a battery backed raid, don't expect more than a few MB/s for "Random write" and "Mixed workload". Also be aware that on disk cache might be active and influence the result. It is however small enough to be saturated by 2-3 threads.

  5. If the disk is bad, replace it before you lose data.
Improve this answer
1
  • You might as well be short of RAM. If the system is doing heavy swapping, the IO is completely trashed.
    – j.hedin
    Apr 24, 2013 at 21:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .