Re: RAC - NOARCHIVELOG to ARCHIVELOG migration

We had an issue with our three-node RAC batch environment. Even at 5 GB redo we were switching about every 15-20 minutes. Since only one of the physical nodes was responsible for the I/O to the archive log destination, it would be overwhelmed by requests to write the large archive logs. Eventually, two of our three instances would shutdown (I think it was 3 or 5 attempts to get their archive logs written before they panicked and shut down). So we wound up creating three disks for archive logs and had each node be responsible for the archive log I/O for that node's instances. This meant that each instance had it's own log_archive_dest.

Kevin Lidh

On 4/25/06, ramick <ramick_at_dotster.com> wrote:
>
> Also, set your log_archive_dest_n to location(s) where RMAN will be able
> to
> find the archived redo if using RMAN to manage these files.
>
> -----Original Message-----
> From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]
> On Behalf Of Andrey Kriushin
> Sent: Tuesday, April 25, 2006 1:54 PM
> To: vikramsingh120_at_gmail.com
> Cc: oracle-l_at_freelists.org
> Subject: Re: RAC - NOARCHIVELOG to ARCHIVELOG migration
>
> And OF COURSE do not forget to backup your database after switching to
> ARCHIVELOG mode.
>
> - Andrey
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

--
http://www.freelists.org/webpage/oracle-l