Re: OPS redo log files in AIX, HACMP

In article <01bc298a$8eca9cc0$81a9bdcc_at_acj607hqt76374>, mks_at_internetmci.com says...
>
>I was wondering what the preferred (or most common) method of storing redo
>log files is (i.e.. raw volumes or file system).
>
>The OPS Concepts and Administration manual implies that redo log files must
>be concurrent and therefore raw. I would rather use file system on each
>node to hold the log files. Anyone else using file system for log files in
>an OPS environment? How about control files?
>
>-matt

On most platforms you're still required to use raw devices for the shared volumes required by OPS. I know that the SP2 still requires them, so i'd guess the HACAMP would as well.

Shared volumes are required for all data, control, and online redo log files with OPS...

Besides, you'll get the best performance from your redo logs by having them on raw devices with a small stripe size.

Steve

-- 
        ,_______________________________________________________,
~~~   ??|                                                       |
c-OO    |  Steve Bourgeois                            
   ) /  |  Oracle Corporation
  -     |  Boston, MA                                           | /__ OO-
==()    |                                                       | |  |o  U
  ||   _|  sbourgeo_at_us.oracle.com, sbourgeo_at_ici.net             | |  |== )=
   \\_//|  http://www.ici.net/cust_pages/sbourgeo/sbourgeo.html |  --  //
    `-' `-------------------------------------------------------'      ~
The statements and opinions expressed here are my own and do not necessarily 
represent those of Oracle Corporation.