on 04-19-2016 3:21 PM
Hello World,
one system ist done, a thousand more to go.
In preparation for an EHP Upgrade of an SCM5.0/LC (APO) system, SUM demanded that we upgrade the old box to SCM SP14 first.
Did that, on both the test and production box.
The test box went through just fine, because and there I did in in many small steps (Basis packs first, and so on).
I also had the latest and greatest SPAM/SAINT and tp/R3trans version installed.
So on the test box all went in w/o a glitch and the dreaded SPAM was even nice enough to interrupt for SPAU and SPDD.
So because I still had a time window open and there was just one guy supposed to be using production anyway, I got cocky and when I installed SP14 on Production I defined one big queue for all packages.
It was like I had been asking for trouble, because this time around SPAM gave me no chance to "reset changes to originals" (=SAP Note corrections), but went straight into kamikaze mode.
SPAM got stuck in the middle (IMPORT_PROPER), complaining "SAP Originals had been changed" and would neither go back nor forth from there.
Finally I had the likewise dumb idea (being pressed by time doesn't make one smarter in such situations) to reset the OCS queue, to be able to reset those changes in SNOTE.
Well that worked, until the system started throwing error message like this one at every turn:
>SAPLSMPI
>A newer version of data type RSMPE_KATR was found than required
The box shortdumps now even when calledn SE38 or SE80, so there is no way how I can (re)activate changed objects.
Is there any way to activate or version change the offender(s) here via tp or another way inside SAP I haven't though of yet?
SGEN still works, but three runs over everything didn't change a thing.
And guess what, right now they are telling me that, "the only one guy using this box", suddely brought a dozen sibblings to work 😞
Hello Donaldo,
Have you tried to set:
abap/pxa=shared,generate
in the shadow instance and restart it?
This will assure that every program is compiled before its first use.
Another possibility (a bold one) is to truncate the repoload table of the shadow instance. This will remove all the loads...
Regards,
Cris
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hah, you were spot on Christiano and many thanks for it!
The only reason why I am keeping this thread open is that now the short dumps are gone, but the system won't connect to the LiveCache no more.
Calling LC10 results in the following error (after being stuck a while trying to retrieve version info):
>DBM_EXECUTE
>dbm_version
>command timeout exceeded
Maybe you can also help me with this one?
Hi Donaldo,
This seems to be a problem related to the livecache.
Would you mind raising a new thread in the livecache community?
Regards,
Cris
User | Count |
---|---|
87 | |
10 | |
10 | |
10 | |
7 | |
6 | |
6 | |
5 | |
5 | |
4 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.