>Are there still problems known or is it generally considered safe to go
>to VSM USEZOSV1R9RULES(NO)? We run 1.10, going to 1.11 soon.
IPL your test system / sysplex with these options in your diag:
Traps Name(
IeaInitArSrb /* Initialize ARs for SRBs */
IeaInitRegsTask /* Initialize GPRs for TCBs */
IgvInitCpool /* Initialize CPOOL GET elements */
IgvInitGetmain /* Initialize GETMAINed storage */
IgvInitFreemain /* Initialize FREEMAINed storage */
IarSt64InitGet /* Initialize gotten 64-bit stg. */
IarCp64InitGet /* Initialize gotten 64-bit cells */
)
In addition, copy igvdgnpp from samplib (?) to your proclib and start that stc.
All of this will basically 'pollute' everything not expressly initialized with
a
pattern of all Foxes or x'4F'. Last time I tried this, we were hit by quite a
few
problems.
On the other hand, if your system comes up and runs without any abend0cx,
you can safely turn on using the new rules. The above diag traps thoroughly
test all code and would simulate errors from the improved rules.
>We ran into a problem with Connect Direct at the 4.6 release level. The
>problem is fixed in Connect Direct 4.7 and 4.8.
Yes, we did too using the above traps. My colleague reported it to Sterling,
and got one fix named T040140. Now that IBM owns Sterling, it should be a lot
easier to get anything fixed that pops up using 'dirty getmain'.
>So far I have found that SHCDMUF (I think this is CA/DATACOM DB)
>needs LCCA and/or PCCA in 24bit, CA/OPERA doesn't like REUSASID(YES),
>and CA/PMO has problems with USEZOSV19RULES(NO).
>I am also still tracking down one or two users of key 8 CSA.
Yes, CA are certainly BIG offenders in this game. Actually, we just ran into a
key8 CSA thing with datacom - you cannot change the XCF MUF group name
without it. Even after IPL they do not come up.
One product that doesn't handle the improved VSM rules is BETA91 and their
BSA. Beta at one point refused to fix the bugs because they said we were
using 'unsupported options'. Which was patently false, as the problem occured
in production, where we certainly won't use the new rules until the *test*
sysplex runs cleanly.
So expect quite a bit of resistance (both from the vendors and from your
colleagues responsible for a buggy product) when you report a bug caused by
the diag traps.
Best regards, Barbara Nitz
----------------------------------------------------------------------
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
