Separate Pool for JDEB7332 Subsystem?

Hi All,

Can anyone think of a reason NOT to put the JDEB7332 Subsystem into a private pool of it's own? This is where JDENET_K's, JDENET_N's and NETWORK and SENTINEL run and they consume much of the CPU resources. If I have them in a separate pool I can monitor and manage resource utilization better. At least that's my hope...

Darrell

AS/400 730 8-way 8gb B7332 SP11.3

Darrel,

I wouldn't think that you'd want to manage all your eggs in the one basket.

It might be best to leave the objects currently in the queue... where they stand... put the other eggs in another basket... This will allow you to give the 'primary JDE' applications - their full bandwidth of the system... and throttle the other applications - based on queue...

Just a thought.

Daniel Bohner
[email protected]
www.existinglight.net
JDE - XE & AS/400
JDE - B7331 & MS SQL 7x

Hi Daniel,

Thanks for the reply. JDEB7332 is already in one basket, *BASE. The problem is that other objects are competing for the same resources as all the JDE stuff. Or am I looking at this wrong.

Darrel,

If I understand - assumption brought in - you want to be able to manage what resources are being used by applications... give some, based on subsystem, greater priority... Correct me if I assumed too much.

You could setup several sbs's... leave the JDE default as it is... that way you don't get yelled at by customer service. Point applictations to the other sbs - and grant appropriate levels at those sbs's... give them more / less than those of competing applications...

This way, JDE doesn't yell at your for changing their defaults, you manage what has more umphf.. based on sbs...

Basically - play with priorities of the SBS... give base more/less than the competitors... or move either one to another/new sbs with differing priorities...

best wishes...!

Daniel Bohner
[email protected]
www.existinglight.net
JDE - XE & AS/400
JDE - B7331 & MS SQL 7x

Actually I think that's a good idea. Especially if you are strapped for
memory. We haven't done that here yet because we're not in full production
yet and the system has a lot of room for growth. You want to consider
putting QSERVER in it's own pool because this is where your interactive One
World batch will run. Good luck trying, it usually involves a little trial
and error in setting up a new pool, meaning getting the correct pool size
and faulting.

BK
Brian Kamps
AS/400 V4R5 B7332 live, XE test



AS/400 V4R5 B7332 live AP/GL, XE/test

Allison,

I have not separated my JDE subsystem into other pools, they run in my
*SHRPOOL1.

What I have done is seperate out my QZDASOINIT jobs to run in their own
pool. We are a mostly TSE shop and these jobs are the ones that service
the connections for the users on the TSE's. We currently have all BSFN's
running LOCAL.

I have also seperated out the batch jobs into seperate subsystems (wrote my
own router), and these run in a seperate poo. The way we are configures
the JDE Kernels do not do a whole lot. As we start moving BSFN's back to
server we will re-visit the need to seperate the memory pools.

One thing I would consider is changing your system values to disable or
severly restrict the autotuning features. The main reason for this is
autotune is biased towards the interactive and we are trying to maximize
server performance.

HTH,

Tom Davidson

Xe Update 1, CO on SQL, ES: AS/400 830 4-way 16G RAM 1T Disk, Two instances
English/Western European & Japanese.



OW 7332 SP 11.3VER, NT 4.0 SP 5, TSE 4.0 SP 4, Metrframe 1.8, CO SQL 7.0