configuration file

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
vijaykumar
Participant
Posts: 228
Joined: Tue Oct 03, 2006 7:08 pm

configuration file

Post by vijaykumar »

Hi,
My company is upgrading from server to parallel, can anyone help me in parallel edition , what is the theory behind assigning no of nodes to a process.

What i think is that if the no of nodes increases then load increases, so how can i determine the exact no of nodes required for a given process.

cheers;
vijay
chulett
Charter Member
Charter Member
Posts: 43085
Joined: Tue Nov 12, 2002 4:34 pm
Location: Denver, CO

Post by chulett »

You forgot the 'regarding'. :wink:
-craig

"You can never have too many knives" -- Logan Nine Fingers
narasimha
Charter Member
Charter Member
Posts: 1236
Joined: Fri Oct 22, 2004 8:59 am
Location: Staten Island, NY

Post by narasimha »

There are many postings relating to this.
Try a search first.
Narasimha Kade

Finding answers is simple, all you need to do is come up with the correct questions.
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

Each job requires one conductor process (on the "conductor node"), one section leader process per node and as many as one player process per stage per node. These are allocated automatically by the parallel engine.
Allocation as described may change due to operator combination and restriction of particular stages to execution in node pools. The former is automatic; the latter is a design decision. Operator combination may be prevented on a per-job or a per-stage basis.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Saama
Premium Member
Premium Member
Posts: 83
Joined: Wed Nov 22, 2006 6:42 pm
Location: Pune
Contact:

Post by Saama »

Hi Gurus,
Thanks Very much.

cheers;
saama
I_Server_Whale
Premium Member
Premium Member
Posts: 1255
Joined: Wed Feb 02, 2005 11:54 am
Location: United States of America

Post by I_Server_Whale »

You can mark the topic as "Resolved".
Anything that won't sell, I don't want to invent. Its sale is proof of utility, and utility is success.
Author: Thomas A. Edison 1847-1931, American Inventor, Entrepreneur, Founder of GE
abc123
Premium Member
Premium Member
Posts: 605
Joined: Fri Aug 25, 2006 8:24 am

Post by abc123 »

Ray, I have a few questions:

1)When we are talking about a "process", we are talking about a operating system process thread, right?
2)If you have a config file with 2 nodes and they look identical other than the name, how do you tell which one is a conductor process, which one is has the section leader and which has a player process?
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

1. No. Just process. As a general rule DataStage does not use threads. (There are, of course, some exceptions, particularly in sorting.)

2. They both have section leader and player processes. The conductor node is usually the first-named node in the default node pool, unless you have specified APT_CONDUCTOR_NODE to place it elsewhere.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
abc123
Premium Member
Premium Member
Posts: 605
Joined: Fri Aug 25, 2006 8:24 am

Post by abc123 »

If there is a configuration file where both nodes are on the same box, does it make a difference as to which one is the conductor?
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

No.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
Post Reply