James, Not sure how you installed all of your Portal. If you installed the entire enchilada on one Linux zSeries then I would expect some extreme overhead.
Recommendation: 1.) If you installed the HTTP sever on the same Portal server from the WAS Portal CD then deinstall the HTTP server and install it on a separate Linux server where it's only mission is to be an HTTP server. 2.) If you installed the nodeagent on the same Portal server also deinstall that and install it on it's own Linux server. 3.) Did you also install DB2 and LDAP as part of the Portal install on the same Linux server ?? If you did then you want to deinstall the LDAP and DB2 from the Portal server and install them on separate Linux servers. You would install the DB2 Connector on the WAS Portal so that it can connect to your Linux sever that is running DB2. The LDAP should be on it's own Linux server and it I believe also requires a DB2 to support LDAP. You can install a DB2 on the LDAP Linus server. Also if you installed the sample applications as part of the Portal install remove those as well. Your WAS Portal startup will be faster and allow you to deploy your own apps. Normally I do not install the sample apps on any Production or UAT Linux servers. I would install them only on a sandbox for applications group to use for training. By moving these main functions to separate servers you can reduce the very high system resource requirements of the Portal server itself. Even with this 680meg of storage for the Portal Server maybe borderline depending on the size and number of apps you are trying to deploy on that Portal server. I hope you have sufficient xstor on zVM that supports your Linux environment. ------------------------------------------------------------------------------------------ James wrote: Linux on 390 Port <LINUX-390@VM.MARIST.EDU> wrote on 02/24/2005 03:12:27 PM: > I have to wonder if the deploy takes a lot of memory for a short > period of time. This would cause other applications to page out, > but by the time you look at memory the deploy has finished using most of it. > Here's the vexing problem: When I have WebSphere running, i get a total memory footprint of approximately 680 megs, give or take 5-10 megs. This is with an HTTP server, deployment manager, node agent and 2 app servers. No swap happens until we try to do the deploy. -------------------------------------------------------------------------------------------------------------------------------------------------- Regards, Terry L. Spaulding IT Specialist IBM Global Services [EMAIL PROTECTED] ---------------------------------------------------------------------- For LINUX-390 subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit http://www.marist.edu/htbin/wlvindex?LINUX-390
