5

Today I get an alert mail from linode.com "Your Linode, Randy, has exceeded the notification threshold (1000) for disk io rate by averaging 1571.81 for the last 2 hours.".

Last time I get alert like that is several month ago, I keeping get alerts, but at the same time my website network traffic never grow. I try everything, but can not location the problem. Finally, I guess the growing log file size is the problem. I use cronolog cut log file, and problem resolved.

Now,I/O rate chart shows I/O rate is keeping growing too, how to I location problem? Is there anything can help me?

My server is Ubuntu 8.04.4 LTS/apache/php/mysql at linode.com.

3 Answers 3

5

You can determine which process is thrashing your disk using the iotop package. http://ubuntuforums.org/showthread.php?t=876738

Also, you can change the I/O rate alert threshold in the Settings tab of the Linode web interface.

2
  • Can I determine which file is write or read?
    – Tinyfool
    Dec 7, 2010 at 23:59
  • Once you see which process is writing, you can use 'sudo lsof | grep process_name' to see which files it has open.
    – slillibri
    Dec 14, 2010 at 23:05
1

Try vmstat and see if there's much bi/bo. I thought it is caused by swap thrashing, due to the small physical memory of your Linode.

1
  • bi=2,bo =4. I don't think so, because seems memory is never run out.
    – Tinyfool
    Dec 7, 2010 at 11:03
-1

1st,you may be find the problem , mysql or apache
using vmstat iostat

2nd, if problem from apache , you can using nginx +fastcgi to replace apache it is a light web server, better than apache

 if problem  from mysql  ,you can modify  my.cnf   

3rd if problem from other app? please paste your log file

1
  • 1
    point is how to use vmstat or iostat find problem from what app? I just can see some global info, not info about a app.
    – Tinyfool
    Dec 7, 2010 at 13:03

You must log in to answer this question.

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