Poor performance on an M3000

Hi

We have an M3000 single physical processor and 8gb of memory running Solaris 10. This system runs two Oracle Databases one on Oracle 9i and One on Oracle 10g.

As soon as the Oracle 10g database starts we see an immediate drop in system performance, for example opening an ssh session can take up to 2 mins. We are also seeing a very heavy processor queue and paging large, what appear to be, paging statistics.

This server has been hardened using JASS and the oracle user has been converted to a role.

Below is an example of a vmstat run - this was done with both databases up but not processing anything:

 
vmstat -S 6 10 
 
 
kthr memory page disk faults cpu 
 
 
r b w swap free si so pi po fr de sr s0 s1 s2 s3 in sy cs us sy id 
9 0 108 20334544 519368 0 0 32 52 59 0 1429 8 6 16 -0 5875 1092 22442 0 7 92 
0 0 34 22395064 790928 0 0 0 0 0 0 0 0 0 0 0 603 552 486 1 0 99 
0 0 34 22394864 790848 0 0 0 0 0 0 0 0 0 0 0 786 1101 711 0 0 99 
0 0 34 22394864 790888 0 0 0 0 0 0 0 0 0 0 0 820 1221 749 0 0 99 
0 0 34 22394864 790888 0 0 0 0 0 0 0 0 0 0 0 838 1315 781 0 0 99 
0 0 34 22394864 790888 0 0 0 0 0 0 0 3 5 10 0 1336 1271 1453 0 0 99 
0 0 34 22394824 790848 0 0 0 0 0 0 0 0 0 0 0 809 1203 734 0 0 100 
0 0 34 22392496 788512 0 0 0 0 0 0 0 0 0 0 0 806 1289 749 0 0 100 
0 0 34 22379680 765104 0 0 0 0 0 0 0 0 0 1 0 830 1311 793 0 0 99 
0 0 34 22364424 747792 0 0 0 0 0 0 0 0 0 0 0 814 1165 764 0 0 99

Just wondering if anyone has experienced the same or a similar problem.

are you on latest patch level? just to rule out known bugs...

Something in /var/adm/messages that looks like a problem?

No sorry nothing in /var/adm/messages worth looking at

Your system looks like undersized in RAM / your applications too demanding.

I agree with Jill, one Oracle DB in 8g ram is pushing it, much less two running different versions of the Oracle DB. I didn't think you could get an m3000 with 8g ram anyway, are you somehow zoning this server to share the resources or has someone replaced the ram with a smaller amount?

As root, run:

And post the results here.