RE: Managing CPU_COUNT for micro-partitioning on AIX

Wow. Which DASDs did they expect you were using? ;)  

And, yes, based on the excellent responses I've had in this thread, I plan to override the _CPU_COUNT. Without checking, I would think that the parameters derived from _CPU_COUNT would be static anyway, but I don't need them changing on me if I should need to bounce the instance(s).  

I don't know if I'll get a chance to read much of the AIX manuals (save for looking up "mpstat"!), but I trust the guys that are calling the shots on the OS.  

Thanks all for your help!  

Rich

From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Mark Brinsmead Sent: Thursday, July 06, 2006 9:52 PM
To: Brandon.Allen_at_oneneck.com
Cc: oracle-l_at_freelists.org
Subject: Re: Managing CPU_COUNT for micro-partitioning on AIX

Cool. I now know a lot more than I did before. Thanks...

Still, in this situation, it would probably be wise to override the CPU count the OS is reporting to the database. The OS is obviously lying. ;-)

Reading the AIX manuals more than the sysadmins do? That sounds all too familiar.

As for not getting "root" passwords, well, that is not uncommon in my experience. As a DBA, I have been given "root" passwords very rarely indeed,
and as a sysadmin, I have provided "root" passwords to others even more rarely. There are matters of responsibility and accountability that need to
be respected, after all. But a sysadmin who disregards (database) tuning advice from the DBAs (or vice versa) is usually ill advised. Usually. ;-)

I do, however, remember times as a sysadmin when I had DBAs adamantly insist that I must places specific data files on specific "cylinders" and
specific "surfaces" of the 500 or so SCSI disks I was managing. After explaining to deaf eafs that neither UNIX nor SCSI really work that way "ignoring" was the only option I had left. Sometimes it *is* a good strategy. ;-)

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