Re: too many archive logs

From: mike <mkid9999_at_gmail.com>
Date: Thu, 19 Mar 2009 05:48:55 -0700 (PDT)
Message-ID: <e5361375-7c94-4284-9031-e5b3bd3ea02e_at_c11g2000yqj.googlegroups.com>



On Mar 19, 7:52 am, mike <mkid9..._at_gmail.com> wrote:
> On Mar 18, 3:08 pm, "Matthias Hoys" <a..._at_spam.com> wrote:
>
>
>
> > "mike" <michael.kidw..._at_hqda.army.mil> wrote in message
>
> >news:ba1dad6c-6e22-43e6-890e-d5eb56db377a_at_v39g2000yqm.googlegroups.com...
>
> > > I'm running 10.2.0.4 on solaris 10 (sparc 64).
>
> > > I'm in archivelog mode with 3 groups of redo logs sized at 200M (at
> > > 100M I was getting checkpoint can't complete messages). I've got
> > > archive_lag_target set to 900. Watching the alert log all seems to be
> > > fine. However, when I go to /flash_recovery_area/SID/archivelog, there
> > > are thousands of archive files, mostly between 10 and 20 MB. For
> > > example, at 6:49  I see six 11MB archives. Some days there are ~ 6000
> > > archive logs. Interestingly, they seem to get progressively smaller.
> > > At midnight the archive logs are over a 100M, by 7am they are down to
> > > 11M with occasional outliers.
>
> > > This seems to be impacting my backups, which are taking 20 hours when
> > > I'd expect them to take less than 2.
>
> > > Thanks in advance.
>
> > Do you have scheduled jobs or other processes which have "alter system
> > switch logfile" in their application code ?
>
> > Matthias
>
> negative.

this is solaris 10. the backup is rman via oem to the local filesystem Received on Thu Mar 19 2009 - 07:48:55 CDT

Original text of this message