RE: [WEB400] V5R3 Apache Performance

["Haas, Matt" <Matt.Haas@xxxxxxxxxxx>]

Ken,

You may want to try posting this to the mailing list at
http://www.ignite400.org.

Since adding more memory to the partition helped, have you looked at the
non-database page faults in WRKSYSSTS? There was an email posted on the
Ignite 400 mail list a few weeks ago from IBM and they like to see under
100/sec for peak performance. If your CGI programs use SQL, you may want
to start the debugger against the CGI service jobs to see if you need to
adjust any of your indexes. I'm pretty sure that there were significant
changes to the database in this release. 

Matt

-----Original Message-----
From: web400-bounces@xxxxxxxxxxxx [mailto:web400-bounces@xxxxxxxxxxxx]
On Behalf Of Ken Lauder
Sent: Tuesday, May 24, 2005 6:47 PM
To: web400@xxxxxxxxxxxx
Subject: [WEB400] V5R3 Apache Performance

We upgraded our server about 8 weeks ago from V5R2 to V5R3.  Everything 
has been fine on 3 of 4 of our partitions but on our 4th partition we
seem 
to have had a distinct decrease in our web performance. 

That partition is setup to run two HTTP servers.  One plain old http and

the other is ssl.  They both function using Mel Rothman's CGIDEV2 code
to 
publish the data.  We are at the latest CUME, HIPER, and GROUP PTF's for

anything and everything pertaining to the system and http.  We have also

had a PMR open with IBM for weeks now with nothing to show.  We've moved

some memory resources over to the *MACHINE of that partition and it
seems 
to help but nothing yet has gotten us back to V5R2 type performance

Has anyone seen this type of performance drop when upgrading to V5R3? 
Could there be something different about how Apache reads/writes data to

the IFS?  Are there any known seizing issues with the upgrade?


Thanks,
Ken Lauder