i can

Dawn May

Dawn May

Bookmark and Share

October 2014

Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4
5 6 7 8 9 10 11
12 13 14 15 16 17 18
19 20 21 22 23 24 25
26 27 28 29 30 31  

« i Can … Benefit from IBM Systems Lab Services and Training | Main | How to Use the Integrated Web Application Server »

September 08, 2010


TrackBack URL for this entry:

Listed below are links to weblogs that reference i Can … Use Subsystems With Many Devices:


Good information, thanks for sharing. What was the last release that the work management manuals where updated for page fault ranges and recommended memory settings?

Hi Travis,
We tend to not document specific performance recommendations any more because there is such a wide variety of applications and workloads that the answer is always "it depends".

General performance tuning information is now in the Information Center under the Performance section (Systems Management -> Performance -> Managing System Performance -> Tuning Performance).

We published a Redbook in November of 2009 that included information on tuning as well as some basic guidelines. The name of this Redbook is "End to End Performance Management on IBM i" and can be found here:

Hi May, interesting the large number of devices in a single sbs, what about the opposite? for example to have 300 active sbs. I have an ISV application that uses a SBSD for a "logical" group of users and now, in a process of consolidation, these groups can be 300. Do you know situations where hundreds of sbs are used succesfully?

Hi Stefano,
As documented by the Maximum capacities section in the Information Center under the work management limits,, the maximum number of active subsystems the system supports is 32,767. Since a subsystem job is just another job, it counts toward the total maximum number of jobs in the system as well. Using subsystems as a way to consolidate workloads on to a single partition is common.

However, as the number of active subsystems increases, there are some operations that may take longer to run as the system has to inspect information about every subsystem. For example, starting and ending subsystems.

Of course, if you start or end a lot of subsystems all at the same time, you can consume a lot of system resources.

The comments to this entry are closed.