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: Size of files. Can you estimate?

Re: Size of files. Can you estimate?

From: Charles Hooper <hooperc2000_at_yahoo.com>
Date: 11 Dec 2006 09:41:09 -0800
Message-ID: <1165858869.123494.301700@16g2000cwy.googlegroups.com>


amerar_at_iwc.net wrote:
> Charles Hooper wrote:
> > amerar_at_iwc.net wrote:
> > > Charles Hooper wrote:
> > > > From your posting, you have only three log files, each roughly 3MB in
> > > > size. Based on your response to Mark, you are switching log files
> > > > roughly every 9 to 30 seconds. Log file switches ideally should take
> > > > place every 20 to 30 minutes. As the system is configured right now, I
> > > > predict that your database will grind to a halt every 9 to 30 seconds
> > > > if the system is set to archive redo mode. You need to significantly
> > > > increase the size and number of online redo logs.
> > > >
> > > > Charles Hooper
> > > > PC Support Specialist
> > > > K&M Machine-Fabricating, Inc.
> > >
> > > Thanks for your suggestion Charles. Based on that, I would think that
> > > not only do the log files need to be increased in size, but we probably
> > > need more than 3 log files.
> > >
> > > I've also looked into the alert log and seen this error:
> > >
> > > Thread 1 cannot allocate new log, sequence 584375 Checkpoint not
> > > complete
> > >
> > > My guess is that this also has something to do with the size and
> > > frequency of the switches, right?
> >
> > Possibly yes. I recall seeing a post in Metalink that sometimes the
> > above will be logged, even when the system is operating properly.
> >
> > Checking the database on my system with the lowest activity (1 user,
> > maybe 10MB of redo generated per day), with forced hourly redo log
> > switches - 16MB per redo log, 4 total redo logs:
> > Thread 1 cannot allocate new log, sequence 6568
> > Private strand flush not complete
> >
> > Looks quite similar to what you posted.
> >
> > Charles Hooper
> > PC Support Specialist
> > K&M Machine-Fabricating, Inc.
>
> I think I'm just going to guess that each archive log file will be
> roughly 6 megs.........and, that after a successful backup or the
> database, the previous archive logs are not needed.
>
> This place is short on disk space as it is..........

Doing the math:
Currently using 3MB in less than 30 seconds Ideally, log switches should occur every 20 to 30 minutes. 3MB * (60 seconds/30 seconds) * 20 Minutes = 120MB as a bare minimum per log file.

Worst case:
Currently using 3MB in 9 seconds:
Ideally, log switches should occur every 20 to 30 minutes. 3MB * (60 seconds/9 seconds) * 20 Minutes = 400MB as a bare minimum per log file.

My calculations are significantly higher than the 6MB that you are suggesting.

Charles Hooper
PC Support Specialist
K&M Machine-Fabricating, Inc. Received on Mon Dec 11 2006 - 11:41:09 CST

Original text of this message

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