SUMMARY: what is my kernel doing?

Thanks to everybody who replied.

Lots of you suggested looking at top, prstat,mpstat, but some told me what
I thought, that I just had too many processes running, or trying to run.

The problem, I believe was tied to the fact that we had to turn off
MTS--an oracle feature that handles multiple oracle connections,instead of
a dedicated process per db connection, there is a set of
dispatchers/servers that can handle 1000's of connections. In case of
having too many processes, the OS was too busy taking proccesses on and
off the run queue, with MTS the number of such processes was greately
diminished and load went back to normal.




On Mon, 6 Oct 2003, NetComrade wrote:

> yesterday I had few sudden spikes in load, which had brought the system
> (solaris 7 running oracle 8.1.7.4 and VxVM) to a crawl (well, ~<12 sec
> response times). This is a 10CPU E4500.
> 
> vmstat showed the following:
> 
> 
>  procs     memory            page            disk          faults      cpu
>  r b w   swap  free  re  mf pi po fr de sr s6 s9 s1 sd   in   sy   cs us
> sy id
>  110 3 0 4798808 375504 0 1928 0 0 0  0  0  0  0  0  0 6289 24099 7686 31
> 69 0
>  102 3 0 4806544 381640 0 1255 6 0 0  0  0  0  0  0  0 6647 27437 8054 33
> 67 0
>  93 2 0 4806408 382664 0 3486 6 1  0  0  0  0  0  0  0 7090 29187 9067 36
> 64 0
>  64 4 0 4806272 382376 0 3811 0 0  0  0  0  0  0  0  0 7106 34630 9294 44
> 56 0
>  59 1 0 4801248 378376 0 2541 0 0  0  0  0  0  2  2  0 7036 29577 9088 37
> 63 0
>  64 4 0 4800496 377256 0 2081 0 0  0  0  0  0  0  0  0 6220 23640 7419 32
> 68 0
>  80 2 0 4803888 380208 0 1841 0 0  0  0  0  0  0  0  0 4966 21493 5448 31
> 69 0
>  87 1 0 4806496 383544 0 2468 0 0  0  0  0  0  0  0  0 5844 22951 6842 29
> 71 0
>  84 4 0 4806656 382400 0 3432 0 1  0  0  0  0  0  0  0 6397 28243 8128 37
> 63 0
>  95 3 0 4801384 377872 0 3440 0 0  0  0  0  0  0  0  0 6843 33454 9052 40
> 60 0
> 
> 
> How can I identify what my kernel is doing the most? Is there are a
> way? The system doesn't seem to be paging much, so I am puzzled.
> Any other performance metrics I should be looking at?
> 
> Thanks,
> will summarize.
> _______________________________________________
> sunmanagers mailing list
> sunmanagers@sunmanagers.org
> http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers