[SunHELP] iowait problem...
Jan Johansson
sunhelp at sunhelp.org
Thu May 3 05:26:45 CDT 2001
On Thu, May 03, 2001 at 01:15:04PM +0300, Devrim Yilmaz (Garanti Teknoloji) wrote:
>Here is my case. This server has lots of iowaits and i want to find the
>reason. If you have any ideas, it makes me happy.
>
>last pid: 1677; load averages: 2.13, 2.16, 2.17 13:01
>77 processes: 76 sleeping, 1 on cpu
>CPU states: 0.0% idle, 2.6% user, 4.8% kernel, 92.6% iowait, 0.0% swap
>Memory: 512M real, 7632K free, 1076M swap in use, 319M swap free
>
> PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
> 578 oracle9 1 60 0 267M 179M sleep 241:53 1.03% oracle
> 572 oracle9 1 60 0 262M 179M sleep 240:11 0.83% oracle
> 574 oracle9 1 60 0 268M 179M sleep 247:19 0.72% oracle
> 622 oracle9 5 59 0 111M 2032K sleep 156:58 0.57% webcached
> 596 oracle9 8 60 0 11M 5216K sleep 67:47 0.34% jre
> 614 oracle9 8 59 0 73M 1440K sleep 140:34 0.32% webcached
> 4250 oracle9 39 30 0 70M 12M sleep 18:54 0.28% java
> 1656 root 1 58 0 2552K 624K cpu 0:01 0.20% top
> 1676 oracle9 5 58 0 13M 4896K sleep 0:00 0.14% httpd
> 1677 oracle9 5 58 0 13M 4656K sleep 0:00 0.09% httpd
> 644 oracle9 4 58 0 14M 7120K sleep 5:10 0.07% httpd
> 1675 oracle9 5 58 0 13M 4656K sleep 0:00 0.07% httpd
> 576 oracle9 1 58 0 263M 138M sleep 255:29 0.07% oracle
> 1665 oracle9 4 58 0 14M 7272K sleep 0:00 0.06% httpd
> 17701 oracle9 4 58 0 14M 7016K sleep 0:39 0.05% httpd
I am no expert on system loadbalancing but I think this top give
you quite a good hint. 1076M swap in use, I think you have io
wait because your machine is swapping (putting memory space on
disk). Solution is to add more RAM, 1024M will stop most of the
disk thrasing me thinks but 1500M would make you ready for the
future.
More information about the SunHELP
mailing list