Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Help me tuning this wait event:log file sync

Re: Help me tuning this wait event:log file sync

From: Howard J. Rogers <dba_at_hjrdba.com>
Date: Tue, 16 Jul 2002 13:17:29 +1000
Message-ID: <ah0371$c4g$1@lust.ihug.co.nz>

"zhuchao" <chao_ping_at_163.com> wrote in message news:d8e45edf.0207151851.5d6d2b06_at_posting.google.com...
> "Howard J. Rogers" <dba_at_hjrdba.com> wrote in message
news:<agm1g5$hsj$1_at_lust.ihug.co.nz>...
> > How big are your redo logs and how big is your log buffer?
> >
> > Regards
> > HJR
> >
> hi, Howard:
> Currently my log_buffer is 2M and redo log file is 256M, it switches
> every 30-60 minutes during normal busy time(9:00 am - 11:00 pm).

That's a respectable switching rate, so your log file sizes shouldn't be an issue.

There's a possibility that increasing the size of your redo log buffer might help. Otherwise, make sure you aren't inducing extraneous checkpoints (in other words, log_checkpoint_timeout should be zero and log_checkpoint_interval should be bigger than your largest redo log). Also make sure you haven't got fast_start_io_target isn't set, and (if this is 9i) make sure fast_start_mttr_target isn't set.

Otherwise, given the application issues you mention below, there's not a lot I can suggest.

Have you got a report.txt or a statspack report handy??

Regards
HJR
> The application is making too frequent commits(now about 40
> commits/second), it is hard for me, as the dba to tune the system. Now
> log file sync is 70% of total wait event. sigh.

> Thanks for your suggestions.
Received on Mon Jul 15 2002 - 22:17:29 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US