Need info regd. Performance Monitoring tolls in PX

amitaguptain · Post by **amitaguptain** » Fri May 04, 2007 3:57 am

Hi,

Do anyone has idea about the performance tools in PX?

and how much virtual memory need to be allocated for the PX project?

Another info is needed about the mount points,i mean how many mount points need to be there for the PX project?

Please post me in case anyone has idea on any of the above points...

Thanks in advance

kduke · Post by **kduke** » Fri May 04, 2007 6:19 am

I need to post a new version which works well with PX jobs. Maybe soon. I think your first step is to put stats in tables so you can measure how it is now against what you may change it to.

All of this is controlled thru config files. Your mount points should match you config file, one per node.

csrazdan · Post by **csrazdan** » Fri May 04, 2007 11:23 am

Following is the guidelines which we follow for setting up PX environemnt based on UNIX mount points:

1. Engine Mount Point - This is the mount point on this DataStage PX engine will be installed. Typical size of this mount point should be 5GB.

2. Project Mount Point - This is the mount point you should use to create your DataStage proejct. Typical size of this mount point should be between 2 GB to 5GB based on number of job in the project.

3. Node Data Mount Point - This is the mount point on which resource and scratch disk files will be created. The number of Node Data Mount Point will depend upon the degree of parallism you want to have in configuration file. For example, if you configuration file has 2 nodes create 2 node data mount points and likewise. The size of node data mount points will depends on the your application data.

4. Project Data Mount Point - This is the mount point on which your can create project data, for example, input files, target files, dataset definition files, reject file, log files etc. The number and size of Project Data Mount Point depends upon your application and how you want to organize it.

Based on how your application grows you can add space to the mount point which actually requires it.

Hope it helps....

ray.wurlod · Post by **ray.wurlod** » Fri May 04, 2007 3:54 pm

When you get version 8.0 there is a particularly wonderful resource estimator tool.

Never use NFS-mounted disks for scratchdisk, and avoid using NFS-mounted disks for resource "disk" (where Data Set and File Set data files are positioned). Avoid those disks being used for extraction and load for other purposes while those phases are being executed.