Re: redo copy latch

sleeps = 1445 (from v$latch) and total_waits =158916 (from v$system_event). So your redo copy latch is only (1445/158916)*100 percent of all the sleeps. This means less than 1 percent. So tuning this latch will not help that much, look for other latches that have a high sleep count as ones that can be tuned.

Also enabling timed_statistics, will show you the wait time in v$system_event and that could help you determing how important 'latch free' is compared to the other wait events. (so the 1 percent could become even less).

I wrote a paper about this method of tuning, let me know if you are interested....

Anjo Kolk

(akolk_at_us.oracle.com)

daud11_at_hotmail.com wrote:

> Anjo Kolk
>
> Do you think you can use the figures below to explain a little bit more,
> please?
>
> NAME, GETS, MISSES, SLEEPS
> redo copy, 1187, 1142 , 1445
>
> EVENT, TOTAL_WAITS, TOTAL_TIMEOUTS, TIME_WAITED
> latch free, 158916, 78842 , 0
>
> regards
> Daud11
>
> In article <364B4210.1958497_at_us.oracle.com>,
> Kristiaan Johan Kolk <akolk_at_us.oracle.com> wrote:
> > My guess is that this potential performance problem is very small compared to
> > other issues that you may face:
> >
> > step 1:
> > - look for 'latch free' in v$system_event, how much of the time_waited is really
> > part of the overall wait time
> >
> > step 2:
> > - look at v$latch to find the redo copy latch and the number of sleeps there
> > are. (compare this to waits for latch free in v$system_event)
> >
> > step 3:
> > - now you can determine how important these latch misses are for the
> > performance.
> >
> > Anjo Kolk
> > (I work for Oracle but I speak for my self).
> >
> > daud11_at_my-dejanews.com wrote:
> >
> > > Hi
> > >
> > > The ratio of MISSES to GETS for redo copy latch is more than 90% for my
> > > oracle database. Does that mean that there is a problem with contention for
> > > the redo copy latch? Should I increase the parameter LOG_SIMULTANEOUS_COPIES
> > > to twice the number of CPU? Currently LOG_SIMULTANEOUS_COPIES = no of CPU
> > > (2).
> > >
> > > Some explanations from you guys out there will be appreciated.
> > >
> > > thanks
> > > Daud11
> > >
> > > -----------== Posted via Deja News, The Discussion Network ==----------
> > > http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
> >
> >
>
> -----------== Posted via Deja News, The Discussion Network ==----------
> http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Received on Sun Nov 15 1998 - 19:25:57 CST