-1

pasted below is the results of running iostat -x 10 10 (a command to look at disk performance):

avg-cpu:  %user   %nice %system %iowait  %steal   %idle
          0.32    0.00    0.17   29.31    0.00   70.19

Device:         rrqm/s   wrqm/s   r/s   w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sda               4.30    74.03  6.49 387.31   131.87  3678.72     9.68   152.44  385.31   2.54  99.92
sda1              0.00     0.00  0.00  0.00     0.00     0.00     0.00     0.00    0.00   0.00   0.00
sda2              4.30    74.03  6.49 387.31   131.87  3678.72     9.68   152.44  385.31   2.54  99.92
dm-0              0.00     0.00 10.49 458.94   125.47  3671.53     8.09   185.92  394.49   2.13  99.92
dm-1              0.00     0.00  1.30  0.00    10.39     0.00     8.00     0.96  537.85 325.38  42.26

How can i find the reason of such disk I/O usage on a server where i install magento. This was running fine before but last few days site almost goes down. But canot find the reason. Please help

pQd
  • 29,561
  • 5
  • 64
  • 106
Mian Zadid Rusdid
  • 1
  • You haven't provided any server specifications, store information, traffic levels or noted any changes that you have recently made. Without this it would be impossible to diagnose the fault or assist you. – Ben Lessani Jun 20 '12 at 21:14

1 Answers1

1

with this info it's hard to say if your server is dying because of swapping or mysql hammering the io subsystem. as first two things:

  • try looking if swap is used / how much
  • try running "show processlist;" in mysql to see if there are long running queries.

in the long run install something like munin or cacti to gather statistics about the system load.

pQd
  • 29,561
  • 5
  • 64
  • 106
  • Thank you for your quick response. The ram of the server is 16GB and only a magento website is running. Output of top command shows around 4GB of swap is using. I tried with SHOW FULL PROCESSLIST but could not figure how to interpret this. – Mian Zadid Rusdid Apr 11 '12 at 19:40
  • take a look at the query duration - if it's more than a second - there's something wrong. your server might be still [ab]using swap despite the fact that you have 16GB of ram... run vmstat 1 for few minutes and see if something is read/written from/to swap; i assume that nothing should be. – pQd Apr 11 '12 at 21:10