Page 1 of 1
Head nodes and Compute nodes in Datastage
Posted: Mon Apr 23, 2012 5:06 am
by satheesh123
Can anyone tell me the differences between head node and compute node used in datastage?
Posted: Mon Apr 23, 2012 7:05 am
by chulett
Really too much like an interview question to be simply answered by anyone here but since Ray has already handled an identical question in the past, I'll link to it:
viewtopic.php?t=126751
Posted: Mon Apr 23, 2012 8:11 am
by PaulVL
Go find the Grid Enablement Toolkit Redbook. It will be helpfull.
In a nutshell:
Head Node = Job coordinator.
Compute Node = worker bee(s)
Don't crunch numbers on your head node.
Don't move data from your head node.
You only have one head node, but you can have a ton of compute nodes. Move work to compute nodes and free up more room on the IP Stack of the head node.
Posted: Mon Apr 23, 2012 6:51 pm
by qt_ky
The head can contain lymph nodes. The computer cannot.
Re: Head nodes and Compute nodes in Datastage
Posted: Mon Apr 23, 2012 10:52 pm
by satheesh123
Thanks for the information provided. It is very useful.