Planet MVS Planet MVS User Experiences ACF2 6.3 Upgrade

ACF2 6.3 Upgrade

 "I find your lack of faith disturbing."
--Darth Vader as he open sourced the plans to the Death Star

Here are my experiences upgrading to CA-ACF2 6.3. Please remember that problems I report here could most likely be from mistakes I made and not from product flaws.

Environment

Problems on the Systems Programmer Test LPAR

NBD means No Big Deal
VBD means Very Big Deal

NBD 2000-02-05 :: During the APPLY of base FMIDs, I got this error on the JES2 exit assembly: 
** ASMA972U Virtual storage exhausted; increase the SIZE option
STATEMENT NUMBER 30176 WAS BEING PROCESSED BY PASS 1 OF THE ASSEMBLER
RESOLUTION: Easy fix, start over and change CAINITE5 to add SIZE(MAX,ABOVE) to the Assembler parm: 
  ADD   UTILITY(ASMA90)                        /* ASSEMBLER ASMA90   */
        NAME(ASMA90)                           /* ASSEMBLER ASMA90   */
        PARM(NOXREF,SIZE(MAX,ABOVE),DECK,NOOBJECT)
 /*     PRINT(ASMPRINT)                        /* LISTINGS DD        */
        RC(04)
        .
NBD 2000-02-05 :: The ACF2 6.3 installation manual (as ones in the past) wants you to put the ACF2 SMP/E proc in the JES concatenation. We have had the ability to have user proclibs via the "JCLLIB" statement since MVS version 4.

RESOLUTION: I never follow their advice. I use JCLLIB.
NBD 2000-02-05 :: The ACF2 manuals are available in Acrobat Reader (PDF) files. The "Product Maintenance Letter" says to run job ACROBAT which isn't in the SAMPJCL library.

RESOLUTION: Run the obvious member named ACFACROB.

Rant:

The ACF2 manuals are in TGZ format (UNIX Gzipped tar file). You run a batch job to unloaded this huge 26 megabyte file from Tape to DASD. I hope you have a good pipe between your MVS system and your PC/Workstation because you need to download it to uncompress it.

After you get the TGZ file uncompressed, you have a directory full of PDF files and you don't have a clue as to which file is which. I created a HTML file called acf263a.html to make it easy to select the ACF2 manual via a web browser. You can download it and put in the same directory as the ACF2 PDF manual files.

CA thoughtfully included logs and directories full of output from the program that created the PDF files in the TGZ file. That's information I need and the TGZ file wasn't nearly big enough anyway. Don't get me wrong, I like getting both Bookmanager (best for Searching) and Acrobat (best for viewing and printing).

After uncompressing the TGZ file on the PC, I then FTP'd the individual PDF files back to MVS into the OE HFS and made them available to be served from the MVS web server.
NBD 2000-02-05 :: Received RC=4 from the HLASM assembling the ACFFDR, message: 
ASMA019W Length of EQUated symbol PSEDUO undefined; default=1
RESOLUTION: Searched StarTCC and found problem #4643: 
When the ACFFDR is assembled using Release 3.0 of the High Level
Assembler, warning message ASMA019W is issued for the following line:

    DS    (L'PSEUDO*1)X

Provided there are no other errors, the assembler return code will be 4
due to this warning, but the ACFFDR can be used without adverse
consequences.

Rant:

This problem was reported on release 6.2 on April 14, 1999. Why haven't they fixed this yet?????
VBD 2000-02-05 :: During the first IPL of ACF2 6.3 we got a User-Friendly Wait State from the IBM WLM function (being the nice guy he is). Here's an edited IPCS run from the Stand Alone Dump.

2000-02-07 :: Problem 8706990 opened with CA via StarTCC. CA tech support responded with this:
The ACFSTCID [default started task ID at my shop] should be used. I asked if the NOUSER logonid was expired because we check both the batch default and started task default logonids to see if they are defined in the logonid database and if they have been canceled or suspended. We didn't check if the logonid had been expired. A fix was written at the ACF2 6.2 release to check for an expired logonid so if the logonid was expired, suspended or canceled it would still be loaded. The batch default logonid is checked first. If it cannot be loaded we issue the ACF7A141 message and get out. The default STC logonid load is bypassed We have to refit the ACF2 6.2 fix (LO57166) that corrects this for ACF2 6.3. I am going to transfer this issue to our level 2 group to have this fix refit for ACF2 6.3. In the mean time you can either define the logonids I gave you earlier or change the default batch logonid so it is not expired and either suspend or cancel it so no one can use it, or you can simply remove it.
RESOLUTION: 2000-02-07 :: CA retrofitted the fix from 6.2 onto 6.3 (APAR TA4786E) that corrects the wait state problem. Same day service! I've IPL'd and the fix works. Yee-haw!
NBD 2000-02-05 :: When I went to open problem 8706990 using CA's StarTCC system via the WWW, I found that the "Open a New Issue" web page uses a form that validates "genlevel" yyyymm. ACF2 6.3 is no longer using genlevels, it is a service pack - in this case "SP01".

RESOLUTION: Entered "200001" to make the validation routine happy. CA hasn't got all of the pieces together for this service pack thingy yet.
VBD 2000-02-07 :: After getting past the WLM wait state problem (corrected by APAR TA4786E), the next IPL was proceeding uneventually until I saw two strange NEW abends I've never seen before:
  1. CA-ENF (genlevel 9901) had a SC78
  2. IBM's PORTMAP, TCP/IP task, had a S378 ABEND
Since PORTMAP creates a SVC dump and TCP/IP automagically restarts PORTMAP, we soon ran out of Auxiliary storage. I stopped TCPIP and cleared all dumps. TSO logons failed so the system was unuseable.

The second IPL had different ABENDs where just about every task in the system crashed. Glad this wasn't on production.

2000-02-09 :: I found that if I IPL'd without the ETFA (ACF2 enhancement product from Eberhard Klemens) that I don't have any problems. I'm FTPing one of the dumps to them now. They said that other customers used ETFA with beta SP01. Hmmmm.

2000-02-11 :: After IPLing the system without ETFA or PORTMAP (see next problem item), I did this:
  • S ETFA
  • Logged off TSO and relogged on.
  • Ran Batch job

    Everything is fine so far...

  • S PORTMAP

    My TSO session and the master console locked up immediately!
RESOLUTION: 2000-02-11 :: None so far - Following up with the ETFA folks who are looking at the dump.
NBD 2000-02-07 :: IBM's PORTMAP TCP/IP started task has socket error and terminates right away everytime it is started.

Hmmmm. PORTMAP works fine under ACF2 6.2.

2000-02-14 :: CA tech support said that some customers got this if the default OMVS user and group aren't set.

Funny thing is the PORTMAP userid has the OMVS stuff so why does the default OMVS user and group come into play????

I found that when the new UNIXOPTS resource was created during the first IPL of ACF2 6.3, ACF2 did NOT copy the default OMVS user and group from the OPTS resource. I made a request to our security administrator to fix UNIXOPTS.

RESOLUTION: Changing the UNIXOPTS to put values in the default OMVS user and group fixed the problem.
Info 2000-02-11 :: While talking to CA technical support, the rep informed me of this information APAR, LI69354, (Solution #: 225) "CA-ACF2 6.3 INSTALLATION CHANGE REGARDING CAKSCINT".

RESOLUTION: Passed information to CICS group.
NBD 2000-02-14 :: Under ACF2 6.3 SP01, I enter this command from the MVS console:
SECTRACE ENABLE,OPENMVS
and get these messages:
CAS2105E SECTRACE OPERAND ERROR FOR KEYWORD: OPENMVS
CAS2106E RSN: UNDEFINED KEYWORD ENCOUNTERED
My Bookmanager version of "CA-ACF2 Systems Programmer Guide" for ACF2 6.3 dated September 1999 says this command should work. This command works on my ACF2 6.2 systems.

RESOLUTION: CA Tech support responded with:
The command is different for 6.3. Look at the Cookbook, page 1-28. The command is now
ST SET,TYPE=OMVS,ID=xxxx,FUNC=yyyy,END
where xxxx is an identifier of your choice, and yyyy should usually be set to ALL.
RTFM (but be sure and pick the right 6.3 manual because not all 6.3 manuals are correct).


Production!

With the above problems solved, we had a succesfull upgrade to ACF2 6.3 and have been real happy with it.

Newsletter

CA has a nice newletter available via email that you can subscribe to at http://support.cai.com/listserv/
This web pages does NOT represent the views of my employer.

Last Updated: 2000-05-27
This web page is © 1997-2001+ by David Alcock. All Rights Reserved.