Greetings. Our current UEX2 submits a batch job to MVS. I successfully tested a slightly modified UEX2 which will Start a MVS started task. The only hiccup I see are the on-line ADASAV SAVEs we run. Here’s a brief overview of the current process…
Job #1: Step #1 - On-line ADASAV SAVE
Step #2 - Batch AOS FORCE PLOGSWITCH P
Step #3 - MVS WAIT for 5 mins
Job #2: Step #1 - Copy current PLOG tape to a vaulted PLOG tape
Step #2 - Uncat the current PLOG tape
Step #3 - Cat the PLOG tape for the next day's processing.
Both the batch ADARES PLCOPY (submitted via Job #1, Step #2) and the Job #2 run the only CLASS=C initiator on the LPAR. This process ensures that ADASAV SAVE tape and the vaulted PLOG tape are in-sync.
I have tried to add a “standalone” ADARES PLCOPY as Job #2, Step #1 (and bump the other steps down). Unfortunately this ADARES PLCOPY terminates with the message “ADARES (PLCOPY): NO DATA TO BE COPIED.” Bummer!!
Again, I’d like to convert our UEX2 exits to Start a MVS started task. Being a started task, I won’t be able to “single thread” these jobs thru that single CLASS=C initiator.
Do you guys have any possible ADABAS or MVS/JCL solutions that will ensure that the ADASAV SAVE tape and the PLOG tape are in-sync? We do have automation software, but I really want to say away from this stuff, as I don’t trust the software (or the people running it)!
Any help you could provide would be greatly appreciated!
from the sequence you describe this sounds right. For the ADARES and the PLOG flags on the PLOG it has no influence if you start this job via batch, via a started task or
What I am really want is to be able to run a “standalone” ADARES PLCOPY batch jobstep which will force a dual disk PLOG switch. This way I can be assured that the prior job (which executed an on-line ADASAV SAVE) and the second job (which ran the standalone ADARES PLCOPY) will keep the ADASAV SAVE tape and ADARES PLCOPYed tape in-sync. That is, I will be sure that the SYN1/2 checkpoint is actually on the ADARES PLCOPY tape.
I want to stay away from issuing something like a ADADBS OPERCOM FEOFPL, because just can’t be absolutely sure that the ADARES PLCOPY started task has completed successfully. And of course if the ADARES PLCOPY started task hasn’t executed successfully, I won’t be able to run a ADASAV RESTONL.
Specifying DISP=OLD is a good idea. I really hate to be the pessimist, but some of the things that could go wrong with the ADARES PLCOPY STC even with DISP=OLD are…
JCL error (tho its been a few years since I have committed one of these)!
VTS, ATL or general tape problems. (And trust me, we’ve had had our share of these).
DDCARD or DDKARTE syntax issues.
Possible NABs (ADABAS RSP 255) or other unforeseen ADABAS response codes.
The ADARES PLCOPYs STCs could be “stacked”. That is, one could already be executing when we submitted the second ADARES PLCOPY. My recollection is that as of ADABAS V71x, the nuc will “reject” the ADADBS OPERCOM FEOFPL.
Again, I hate to be the pessimist but the ADASAV SAVE and ADARES PLCOPY tapes really need to be in-sync.
If you force a plog switch (in Job 1 step 2) Adabas nucleus will call UEX2 and UEX2 would normally submit a PLCOPY job that empties the dual plog nucleus just closed. If you later do another PLCOPY there will be nothing to copy.
If your objective are to have SYN1 and SYN2 on same plog tape you should issue the plog switch (e.g. opercom FEOFPL) just before you start your online save. But the dual plog may run full before the save finishes and SYN1 and SYN2 will in this case not be on the same plog tape.