Path for ResourceDisk and ScratchDisk in Config 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
nvalia
Premium Member
Premium Member
Posts: 180
Joined: Thu May 26, 2005 6:44 am

Path for ResourceDisk and ScratchDisk in Config File

Post by nvalia »

Hi,

We currently are starting a New Project and have all Tiers on the same server.

1. Can we have the Resource Disk and Resource ScratchDisk on a separate disk (on the same server) other than the Install Drive by altering the entries in the config file. The install drive is C: where as we want to use F: for temp

2. Any Performance Issues or any other impacts due to this config change

e.g. default is
resource disk "C:/IBM/InformationServer/Server/Datasets" {pools ""}
resource scratchdisk "C:/IBM/InformationServer/Server/Scratch" {pools ""}

Can we have it as

resource disk "F:/Datasets" {pools ""}
resource scratchdisk "F:/Scratch" {pools ""}

Thanks,
NV
PaulVL
Premium Member
Premium Member
Posts: 1315
Joined: Fri Dec 17, 2010 4:36 pm

Post by PaulVL »

Have you tried it?

Given the nature of the question, I would recommend reading the documentation. I do not mean any offence by that. I just feel that you need to get a good grasp of the tool and the documentation can aid you. Experimentation in DEV will also work wonders.


From a technical standpoint, if you have 3 physical drives: C for your engine binaries, F for your DataSets resource disk and E for your Scratch disk, you could take advantage of the hard drive cache for each of those devices and that helps speed things up. You will also be faster because you have more platters spinning (ya old school non SSD statement there) at the same time servicing different files. That is a speed increase.


Rule of thumb:

Scratch: get a fast local disk that is large enough to handle your data volume + elbow room. (SSD would be great)

engine binary mount: Seperate this mount from your customer data drops. You do not want to crash the sapplication because you put to much data in your jobs.

Resouce disk: Potential BIG mount, might need to be shared with another server if you ever cluster/grid your environment.
nvalia
Premium Member
Premium Member
Posts: 180
Joined: Thu May 26, 2005 6:44 am

Post by nvalia »

Thanks Paul.

I am going to test it in Dev but wanted to see if someone has already tested/used this scenario and if so it helps to get some insight on that
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Post by ray.wurlod »

As well as all that I would, if possible, put resource disk and resource scratchdisk on separate drives. They are accessed quite differently. Scratch disk tends to have a more random pattern of access, where as DataSet segment files are effectively read sequentially (each).
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