-2

My server is using two SATA disk Raid 1. I test disk using dd

#!/bin/bash
i=0
MAX_LOOP=10
while [ $i -lt $MAX_LOOP ]
do
        dd if=/dev/zero of=/data/output bs=1G count=10
        ((i++))
done

The write speed is very good about 310MB/s. But IO wait increases during test process then CPU load average also grows up. I use iotop to get the processes who make io wait. Those processes are kmirrord, jbd2/dm-5-8 and dd.

I don't know what cause of the problem. Do you have any ideas?

Improve this question
2
  • 4
    You are making the disks work as hard as possible to copy garbage, and you are surprised that other processes must wait?
    – Michael Hampton
    Dec 23, 2014 at 4:10
  • Actually, three days ago, when I rsynced data from other server to this server. Sometimes, munin system notified a warning disk latency. Total data is rsynced about 20MB. rsync job is run every minute.
    – Luke Nguyen
    Dec 23, 2014 at 4:29

1 Answer 1

Reset to default
0

This is most probably because you force your disks to work as fast as they can. But it may also happen when your server is overloaded, and disk test is only adding even more load. Try to test in offline mode, with all your services turned off.

This might be also an indication of possible hard disk failure in future.

Improve this answer
1
  • Thank for your time. The problem disappeared after I tuned cron job scheduler. This job run one time each hour instead of one time each minute (I hope you understand because I am not sure about English vocabulary)
    – Luke Nguyen
    Jan 3, 2015 at 4:26

You must log in to answer this question.

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