View previous topic :: View next topic
|
Author |
Message |
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
I have PLI DB2 pgm, which updated the table. The problem is : if pgm failed, there is no DB2 rollback. I don't use commit comand in my pgm. I've forced abend in pgm right after first update in test. The result - DB2 commit occured, row was updated. Please advice, what may cause this problem.
Thanks |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
kind of hard to believe!
not likely to be solved remotely ...
too many questions to ask
too much data to be collected posted and analyzed
the road to go is Your support! |
|
Back to top |
|
|
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
I agree, it is like I doubt that 2*2=4 , but it is test result and I have it. Test was simple - abend right after first update and ...ops. I can provide any aditional info, just want to believe again, that 2+2 = 4 |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
since You are sure, no reason for anybody to waste time on a DB2 error
odd that You are the first one to notice
time to get Your support involved and escalate the issue to IBM |
|
Back to top |
|
|
don.leahy
Active Member
Joined: 06 Jul 2010 Posts: 765 Location: Whitby, ON, Canada
|
|
|
|
Are you 100% sure that the program really abended? i.e. that no other process intercepted the abend request and turned it into something else? |
|
Back to top |
|
|
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
I've usedh STOP comand ( pgm is PLI) after sql update... it abended with 1000 and commit the update |
|
Back to top |
|
|
Akatsukami
Global Moderator
Joined: 03 Oct 2009 Posts: 1788 Location: Bloomington, IL
|
|
|
|
anatol wrote: |
I've usedh STOP comand ( pgm is PLI) after sql update... it abended with 1000 and commit the update |
That is not an abend! |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
as I imagined You wasted everybodys time
I just run a quick test with the hello world
Code: |
000049 //HELLO1 EXEC IBMZCPL
000050 //C.SYSIN DD *
000051 HELLO: PROCEDURE OPTIONS(MAIN);
000052 PUT SKIP LIST ("HELLO, WORLD FROM PL/I !");
000053 END HELLO;
000054 //HELLO2 EXEC IBMZCPL
000055 //C.SYSIN DD *
000056 HELLO: PROCEDURE OPTIONS(MAIN);
000057 PUT SKIP LIST ("HELLO, WORLD FROM PL/I !");
000058 STOP ;
000059 END HELLO;
....
IEF142I ENRICO1 GO HELLO1 - STEP WAS EXECUTED - COND CODE 0000
....
IEF142I ENRICO1 GO HELLO2 - STEP WAS EXECUTED - COND CODE 1000
|
hello1 ended with a cond code 0
hello2 ended with a cond code 1000
it would be wise, to review Your perception of Your skill level
from Your profile
Quote: |
Mainframe Skills: good |
too high for somebody who cannot distinguish a non zero return code from an abend
let' s not talk about Your self confidence |
|
Back to top |
|
|
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
sorry, Enrico... you suggest ME not wasting anybody time ... and then question MVS knowled level of Akatsukami - I hope |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
I suggested that You wasted everybody' s time, yes very true
but... where in my post, by grace, did I question Akatsukami' s competence
am I questioning Your attitude ? certainly yes
the quick test I run was intended to show You the difference between return/cond code and abend |
|
Back to top |
|
|
Nic Clouston
Global Moderator
Joined: 10 May 2007 Posts: 2455 Location: Hampshire, UK
|
|
|
|
If you want to force an abend then signal error. |
|
Back to top |
|
|
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
ok, now you convinced me ... I've changed the profile lol |
|
Back to top |
|
|
don.leahy
Active Member
Joined: 06 Jul 2010 Posts: 765 Location: Whitby, ON, Canada
|
|
|
|
Akatsukami wrote: |
anatol wrote: |
I've usedh STOP comand ( pgm is PLI) after sql update... it abended with 1000 and commit the update |
That is not an abend! |
LOL! I knew it had to be something like that. I am not a PLI guy or I would have asked the TS a more specific question. |
|
Back to top |
|
|
anatol
Active User
Joined: 20 May 2010 Posts: 121 Location: canada
|
|
|
|
btw... signal error right after insert ( ON error block has only ON ERROR SYSTEM) :
Current Condition:
IBM0280S The ERROR condition was raised by a SIGNAL statement.
checking table - row inserted ... no need to answer this topic, just for fyi before close the topic.[/code] |
|
Back to top |
|
|
|