In this case so high sys time is strange.
According your vmstat output, cs (Context switch) very low, and problem
not in task switch.
please try to play with filesystem options, none or directio. I see in
vmstat wa (IO wait) is always near 0, but
can see blocked process.

Also I see up to 19M output io. which elevator used?

About listener. How many sessions per minute?

Give me please output of sqlplus show sga and cat /proc/meminfo

Andrew Edunov wrote:
> I'm using IBM ds400 fc disk system. Filesystem io options set to "async",
> ext3. listener.log looks fine.
>
>  
>> From my point of view, you use data files on filesystem and/or software
>> raid. And application create new sessions very often.
>> If you use software raid, this cpu utilization are normal :) espesially
>> if you use raid 5
>>
>> Else, please try to setup filesystem io options to DIRECTIO, mount
>> filesystem with noattime.
>> If you filesystem does not support direct io, please think about
>> ASM/raw/ocfs
>>
>> btw, did you check listener.log?
>>  
>
>
>
>  

--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting