There are a couple of issues that were actually going on. In 1.7 the INTRDR processing is done in the Address Space allocating the Reader. It is possible that an Address Space that is not optimally qualified in the WLM Service Policy, may not get access to the Processor as much as you'd like. The INTRDR processing may become a victim to this. Resist the tempation to put CA7 into SYSSTC. Rather, give it a respectable goal in something like STCHI. Now, if you have indiscriminately placed many STCs in SYSSTC, STCHI may not be good for CA7, and then I'd say put it into SYSSTC until you clean up the Service Policy.
Accessing the JES2 Checkpoint may cause challenges (see next paragraph). Creating the "Hot Potato" effect or having a MAS member Hold onto the checkpoint too long may have impact on CA7. The MASDEF specs for HOLD and Dormancy may need to be updated. See the JES2 Availability and Performance Redbook. It helps you determine the correct specs based on the number of members in the MAS, and what the members are doing. The member that has the CA7 address space (that is the one submitting the Jobs) should have a higher Hold value than the other members. The systems with not much JES2 activity will benefit adjusting the MINDORM and MAXDORM. The real issue: when you migrate from a previous version of CA7 to the current one, you need to be sure that your Configuration specifications have been updated on the new version. A specification in CA7 tells CA7 to retrieve and track the Jobnumber for every Job submitted. If you are submitting large numbers of Jobs, this will put quite a strain on the JES2 Checkpoint. In previous versions, we had customers turn this option OFF. When you forget to do the same thing in the current version, you go back to putting the strain on the Checkpoint. When we had customers experiencing this problem, turn OFF that option, the apparent slowdown went away. Moving CA7 to SYSSTC was done at the same time; but given you all know how I feel about too much in SYSSTC, turning of the Track Jes2 Jobnumber option is the real fix. If you have further questions (CA7 performance, JES2 MASDEF specs, or WLM overhauls), feel free to contact me offline at [EMAIL PROTECTED] Thanks, Chris, for keeping me honest. I'm on the road, too; and I actually could have answered this last week. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
