Re: Tuning 'log file sync'

"Mladen Gogala" <gogala_at_sbcglobal.net> wrote in message news:pan.2005.09.05.16.40.01.103746_at_sbcglobal.net...
>>
>> I don't follow your logic here. What if the transaction's data is at
>> the end of a large buffer?
>>
>> -Mark Bole
>
> Actually, it was a word of mouth. My understanding is that log_buffer is
> written in the several I/O requests and that transaction is notified as
> soon as its commit record is written to the disk, not necessarily waiting
> for the whole buffer to be written down. When I come to think of it, 10.2
> event doesn't prove it as it would need a waiting address, too.
>
> --
> http://www.mgogala.com
>

There is an interesting oddity I am trying to work out. I have a case where the log file syncs are (occasionally) much longer than expected - but only in very short bursts, perhaps a few seconds every 30 minutes or so.

The oddity is that I can see lots of sessions (in one case 93) waiting for the same log buffer to be written. The problem is that it isn't possible for all those sessions to have redo information in the same buffer - so how do they find out that they are waiting for that buffer to be written, it seems to be an unnecessary use of resources to tell them.

-- 
Regards

Jonathan Lewis

Now waiting on the publishers:    Cost Based Oracle - Volume 1

http://www.jlcomp.demon.co.uk/faq/ind_faq.html
The Co-operative Oracle Users' FAQ

http://www.jlcomp.demon.co.uk/appearances.html
Public Appearances - schedule updated 8th July 2005