CA 7 Release 11.3 to 12.1 Upgrade questions not mentioned in the upgrade notes.
Article ID: 205192
Updated On:
Products
Issue/Introduction
Questions regarding the upgrade process which weren't mentioned in the upgrade notes:
1. Is the a SYSMDUMP DDNAME absolutely necessary to include in the CA7 JCL? The reason I ask this question is because I'm concerned that if CA7 captures a dump during execution then CA7 may ABEND with x37 due to there not being enough disk space at the time during the ABEND. By removing the SYSMDUMP DDNAME from the CA7 job if an SVCDUMP needs to be captured then why not let the dump be captured to a system dump dataset rather than a SYSMDUMP dataset and risk CA7 ABENDing with a X37 ABEND?
2. During the upgrade from R11.3 to R12.1 we will have the R11.3 tracking environment support programs (CAIRIM, SVC, SMF exits) as active in the system. Can we continue using the R11.3 tracking environment after upgrading from R11.3 to R12.1 with the view to installing the R12.1 tracking environment only after we know we are NOT going to fallback from R12.1 to R11.3?
3. What type of database backup needs to be performed to reclaim URI's belonging to deleted rows? SEQ=PHYSICAL or SEQ=NATIVE?
4. To perform the upgrade from R11.3 to R12.1 what UID should I be using? UID = 255 ?
5. At what point during the upgrade do I stop CA7NCF and ICOM? Do I stop these tasks only AFTER all jobs in the ACTIVE queue have successfully ended?
Environment
Release : 12.1
Component : CA-7
Resolution
(1). Is the a SYSMDUMP DDNAME absolutely necessary to include in the CA7 JCL? The reason I ask this question is because I'm concerned that if CA7 captures a dump during execution then CA7 may ABEND with x37 due to there not being enough disk space at the time during the ABEND. By removing the SYSMDUMP DDNAME from the CA7 job if an SVCDUMP needs to be captured then why not let the dump be captured to a system dump dataset rather than a SYSMDUMP dataset and risk CA7 ABENDing with a X37 ABEND?
Answer:
The SYSMDUMP dataset is only used for dumps related to a CA7ONL abend, i.e. CA7ONL is terminating abnormally. SYSMDUMP is required because SYSUDUMP does not include 64-bit storage, which release 12.1 uses.
If CA 7 generates a dump but is not abending, the dump will be written to a system dump dataset - not SYSMDUMP.
(2). During the upgrade from R11.3 to R12.1 we will have the R11.3 tracking environment support programs (CAIRIM, SVC, SMF exits) as active in the system. Can we continue using the R11.3 tracking environment after upgrading from R11.3 to R12.1 with the view to installing the R12.1 tracking environment only after we know we are NOT going to fallback from R12.1 to R11.3?
Answer:
It will probably track smf records ok, so jobs should track in ca7 as they should, but I would upgrade CAS9 to 12.1 asap. We prefer to be downward compatible (12.1 CAS9 with 11.3 CA7,ICOM) versus the other way around. But if you have to go back which we don’t recommend you need to follow all steps. But CAS9 would be ok as is in that case.
You can start CA 7 r12.1 with the current 11.3 CAS9 loaded environment without any issues. But I recommend you update it for your next IPL when CA7 r12.1 goes into Production.
(3). What type of database backup needs to be performed to reclaim URI's belonging to deleted rows? SEQ=PHYSICAL or SEQ=NATIVE?
Answer:
The 'Reorganize Your Database' section. A SEQ=NATIVE backup is required.
(4). To perform the upgrade from R11.3 to R12.1 what UID should I be using? UID = 255 ?
Answer:
UID is for specific things. Like who can do a demand, or issue a ljob and see that job that has a specific UID. But if you have UID 255 then that says you can access any job no matter the UID on the job screen. No direct correlation to the upgrade.
The only commands that are affected by a UID are Demand, so if a job has a UID of anything other than 0 then you need a uid that matches or 255(which is the highest UID and says you can demand any job. The other 2 commands affected are LJOB if it includes a job that has a UID that you don’t have one to one relationship with or you arent UID 255 then you won’t see it in the output. And the last thing that it affects is the job screen. So if you try to list a job that you don’t have that same relationship or you aren’t 255 then you can see the job screen for that job.
(5). At what point during the upgrade do I stop CA7NCF and ICOM? Do I stop these tasks only AFTER all jobs in the ACTIVE queue have successfully ended?
Answer:
You normally bring down all tasks at the same time and bring them back up one at a time. As long as you don’t IPL the Lpar that CA 7,ICOM run on then you won’t lose any SMF records. So no tracking lost.
Feedback
