From: "Craig A. Sharp" <csharp2a@earthlink.net>
>
>I am the Unix Administrator for my company. We currently run an
>Informix database server on an HP9000 system running hpux 10.2.
>I have been asked to install the patch that provides the NOAGE parameter
>for Informix.
>
>Currently, the system is severely processor bound with run queues as
>high as 5.0 with an average of 3.0 during a normal day.
>We are also experiencing memory pressure and IO bottlenecks. We do know
>that the system is severely undersized.
>
>My understanding is that the NOAGE parameter is supposed to prevent
>processes like oninit from loosing priority due to standard system
>priority aging. The thought is that the NOAGE parameter may
>significantly increase the system performance.
>
>HP has suggested that the patch is a dangerous patch to install and runs
>the risk of crashing the system and requiring a full reload of the OS.
>The patch is a large patch and is said to replace many libraries on the
>system.
>
>I am looking for information on what level of performance increase is
>gained by setting the NOAGE parameter. Is there a significant increase
>in processing speed or throughput when this parameter is used. Please
>provide real world examples if possible.
>
>My final objective is to evaluate the gains in performance vs. patch
>risk factors when implementing the parameter on an HA system.
You'd be better off shipping us your onconfig file, output from onstat -d,
onstat -g iov, onstat -g iof, onstat -p, onstat -D and onstat -P | tail -5
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com