Re: Fwd: AIX Concurrent I/O for Oracle archive logs
- From: David Roberts <big.dave.roberts@xxxxxxxxxxxxxx>
- To: rjoralist@xxxxxxxxxxxxxxxxxxxxx
- Date: Tue, 25 May 2010 22:28:02 +0100
For clarity, my theory is that Oracle is being overly cautious in
its recommendation to avoid specific file systems on the basis that the
difference in the block size of the file system and the difference in the
assumed block size for redo and archive logs may inhibit the performance of
Oracle, hence the warning, while the substantially improved performance of
the underlying file system makes that warning moot.
Dave
On Tue, May 25, 2010 at 10:17 PM, Rich Jesse <
rjoralist@xxxxxxxxxxxxxxxxxxxxx> wrote:
> > From: David Roberts <big.dave.roberts@xxxxxxxxxxxxxx>
> >
> > With the appearance of Advanced format hard drives, Oracle has indicated
> > that they want to know the underlying disk physical file format for ASM
> file
> > systems and redo logs.
>
> However, the OP's asking specifically about *archived* redos. If SETALL is
> used, redo log files are automagially attempted to be opened CIO on AIX.
> At
> least on my 5.3 system they are as shown the "lsof +fg" command the Brandon
> posted.
>
> My current redo generation and RMAN backup patterns don't generally give me
> enough of a window to verify this for archived redos however, nor FRA files
> for that matter. And as the oracle account here doesn't have access to
> /dev/mem nor /dev/kmem and I don't have access to root, I'm afraid I'm not
> going to be of much more help.
>
> Perhaps a shell script looping the lsof running in a semi-tight loop (with
> a
> 1-second sleep perhaps?) while generating excess redo would show something
> fun?
>
> Rich
>
> --
> //www.freelists.org/webpage/oracle-l
>
>
>
Other related posts:
