Re: Performence on Oracle instance running on Sun Solaris/Sparc

"andrew_webby at hotmail" <spam_at_no.thanks.com> wrote in message news:<990610214.15678.0.nnrp-13.c30bdde2_at_news.demon.co.uk>...
> Tommy
>
> I think you should take a utlbstat/utlestat snapshot and post the results. I
> also think it is most likely NOT a memory issue because of lack of 'sr'. It
> doesn't matter really that SYSTEM tablespace is almost full. When that
> becomes an issue, you will get an error and not a performance problem.
>
> I notice that your redo logs are on the same disk as oracle binaries - in
> which case, if you are having connections coming in (client's attaching)
> during your work, then there will be sever contention on that disk which
> will result in waits...
>
> A proper utlbstat report will show buffer cache/library cache etc info which
> will be more useful.
>
> Andrew
>
> "Tommy Svensson" <tkman23_at_hotmail.com> wrote in message
> news:ebf7c6db.0105221323.3860f883_at_posting.google.com...
> > Hi!
> >
> > We have huge and strange performance problem on one of our oracle
> > instance
> > running our ERP software on Sun Solaris Servers and OS.
> > We have done a lot of tuning and testing to found the performance
> > problem but
> > without any luck :-(
> <SNIP>

One or two other INITORA parms:

sort_area_size? Since these are reports, I presume lots of group by and order by stuff.

multi-block-read-count? Large disk io if small multi block reads

async io/multiple dbwr's ?

db_block_size?

Cost based or Rule Based optimization--when was analyze last run if cost? Received on Wed May 23 2001 - 10:34:41 CDT