Stephen.
I have nothing scientific to base my assumption on. I'm monitoring top or
topas which is included in AIX 4.3.3. I suspect we are CPU bound because of
the decreased application performance and markedly less CPU idle time. The
tivolivg volume group is mirrored across 3 disks and the rootvg volume group
is mirrored across 2 disks. I/O wait is not a problem. I also suspect the
CPU because of the effect adding a SmartSet had on netmon. I also had
problems with SNMP Configuration using the SmartSet selection. nvcold will
consume 25% of the CPU (1 CPU) alone.
Application performance has suffered over the last several months. The only
major change has been increased retries (from 3 to 5). The number of objects
in the database has remained relatively constant (26,xxx). The number of
users has increased slightly to 23 - 27. The paging space usage seems to
indicate a memory leak. PS increases rapidly (and continues to increase)
after 22 users open a map. I hope the 6.0.3 nvlockd fix will take care of
that.
Thanks.
Ray.
-----Original Message-----
From: owner-nv-l@tkg.com [mailto:owner-nv-l@tkg.com]On Behalf Of Stephen
Hochstetler
Sent: Monday, December 17, 2001 3:27 PM
To: IBM NetView Discussion
Subject: Re: [NV-L] Capacity Planning ....
Ray,
Ok, I will bite. Why do you think your machine is CPU bound? What
performance tools do you have installed? You have a LOT of horsepower
there.
Are you running software other than NetView on this machine?
Kind regards,
Stephen Hochstetler shochste@us.ibm.com
International Technical Support Organization - Austin
Office - 512-436-8564 FAX - 512-436-8701
ITSO redbooks at http://www.redbooks.ibm.com
_________________________________________________________________________
|