Re: RAC on QFS

From: hpuxrac <johnbhurley_at_sbcglobal.net>
Date: 27 Jan 2006 05:55:57 -0800
Message-ID: <1138370157.955335.86260@g43g2000cwa.googlegroups.com>

A database that will be growing rapidly makes a stronger case for at least examining the ASM possible implementation in more detail.

The drawback of RAW is and always will be it takes more attention and time to setup and manage RAW devices. It is still something to consider with a database that will be increasing in size but planning ahead and not getting behind takes more resources.

The size of raw partitions and what people choose to use varies widely.  At the moment, we are using 2, 5, and 10 gig sizes for the data files.

We cross mount the archive logs. Each instance ( 2 node config ) has a local directory for DEST_1 and a remote NFS mounted one for DEST_2.

You cannot however always rely on DEST_2 as being available. I use an exit to do an something like this ...

   case

      when v_the_instance = 'xxxxx1'  then
	     execute immediate 'alter system set log_archive_dest_2 =

''location=/u02/zzz/archive/xxxx02 optional'' scope=memory sid=''xxxxx1''';

Along with some scripting that periodically checks to see if any of the ( copies ) that should be somewhere are not. Received on Fri Jan 27 2006 - 07:55:57 CST