confusion and request for help setup Windows UNIX datastage

dsex100 · Post by **dsex100** » Wed Apr 01, 2009 8:32 pm

Hello,

I HAVE THE FOLLOWING SETUP:

DataStage Enterprise Edition 7.5x2 installed running on Windows Server 2003 Enterprise Edition SP2 (both client and server components)

single processor machine
----

I'm only writing to text files right now to learn the Enterprise Edition.

I will install some databases, starting with Oracle or SQL Server, not sure yet.

I WANT TO SETUP THE FOLLOWING:

I have another machine with a dual-core processor.

I want to setup DataStage on that other machine in order to use the parallel configuration capabilities.

I want to install UNIX on that other machine in order to run the jobs on UNIX.

So, I'm in the process of installing CentOS-3.9 which is the redistributable of Red Hat Linux Enterprise Server 3 (formerly known as Red Hat Linux Advanced Server 3 as in the datastage documentation)

I had some problems with the installation of CentOS-3.9 not detecting my monitor because of drivers so it was recommended to me to install CentOS-5 in order to not have a hardware driver problem.

(This would be the same as Red Hat Linux Enterprise Server 5)

So well, I would like to know if anyone has run datastage on CentOS-5 or Red Hat Linux Enterprise Server 5?

I'm a little confused as to how I will be able to or how I can work with Datastage, setting up between Windows and UNIX between 2 machines.

In the documentation and as I have tested, you can run parallel jobs on Windows Server 2003..this is fine.

Will I be able to kick off parallel jobs from the Windows Server 2003 over my network connection where the jobs are actually located on my UNIX machine?

If so, will the jobs appear in Datastage Director even though they actually reside on the UNIX machine?

I was reading in the Datastage documenation that you can install Datastage on a UNIX machine and use the Datastage client components on Windows XP or 2000.

So, would this also be possible from the Windows Server 2003 "client?"

Something else that made me wonder- in one of the videos in the DSXCHANGE LEARNING CENTER called "WINDOWS: Start dssh" by Ken Bland he connects to a Telnet Session on Windows Server 2003.

Is he connecting to a UNIX machine where the Datastage server components are installed or is he using a UNIX server on Windows??

If it would be possible to use a UNIX server environment that's already on Windows that would be amazing.

Well as you might notice, I'm sort of confused and need to figure this out before I can figure out how to setup database connectivity between these two environments.

chulett · Post by **chulett** » Wed Apr 01, 2009 11:33 pm

A couple of things... one is that you don't need multiple processors to "use the parallel configuration capabilities". Node <> cpu.

The other is you have a Windows version of the tool, so no UNIX anything for you. Not without a license purchase for your flavor of UNIX, that is.

dsex100 · Post by **dsex100** » Thu Apr 02, 2009 12:00 am

hmm so node<>cpu ok..

would you say that parallel jobs are faster to run on the server even if there's only 1 processor compared to datastage server jobs?

and what was I thinking it's true I didn't even think that my version of datastage would only install on windows! duh!

dsex100 · Post by **dsex100** » Thu Apr 02, 2009 12:10 am

thank you for your reply , it really set me straight and on the right track

wow

ray.wurlod · Post by **ray.wurlod** » Thu Apr 02, 2009 1:13 am

dsex100 wrote:would you say that parallel jobs are faster to run on the server even if there's only 1 processor compared to datastage server jobs?

Probably not for small to medium volumes, because the startup overhead of a parallel job is significant compared to that of a server job.
Possibly not for large volumes, because the machine itself will be more likely to be overloaded, not by the total volume necessarily but by the extra processes' demands for resources.

dsex100 · Post by **dsex100** » Thu Apr 02, 2009 9:24 am

thank you for your help

i went through the readme file for my version of datastage and I found 2 locations where it specifically stated or answered my question if it would run on UNIX,

1- Build Platform: WINNT
2- Platform - It states that this product is only intended for Windows
clients- windows xp windows 2000
server windows server 2003 standard or enterprise

i will mark this as resolved and continue my questions in another post

kcbland · Post by **kcbland** » Thu Apr 02, 2009 10:07 am

Just to be clear - you purchase either a Windoze or a Unix version of DataStage. The Clients are always on Windoze, the Server is platform selectable. Everything resides on the Server side. If you write your jobs to be transportable, they can run in either Server environment. Most companies have a single consistent Server platform and don't care about transportability.

As for parallel performance on a single cpu, the answer is usually quite simple. The more tasks you have a single cpu doing the fewer amount of cycles it will spend on those tasks. There's also more overhead of task coordination. The more cpus you have the more you can spread tasks across cpus and the thread coordination overhead becomes distributed.

kcbland · Post by **kcbland** » Thu Apr 02, 2009 10:09 am

In case I wasn't absolutely clear - you buy a Windoze license and it only installs on Windoze.

chulett · Post by **chulett** » Thu Apr 02, 2009 2:41 pm

And if that still wasn't clear - crucifix combined with leg triangle!

JRodriguez · Post by **JRodriguez** » Thu Apr 02, 2009 3:11 pm

Nowdays - IIS 8.xx - the licenses are for all supported OS .... IBM allows you to download the software for any supported OS. The xml license file is the same for all