It seems that when I run jzos a number of times, perhaps with various errors, I no longer get any valid error msgs,
even in response to a deliberate class not found condition.
The SYSPRINT dd is empty and the SYSOUT DD has info and 'T' msgs (LOGLVL=+T) but not 1 thing about the error which should have been class not found in this case.
But when I started running this particular job i did get legitimate class not found and other errors and abends, so maybe it is something with the JVM not getting reset or some other condition is not getting cleared out after a number of runs and errors.
Are there any known problems about this, or did I just mess something up in my job stream?
jzos stops giving error msgs
Re: jzos stops giving error msgs
Please post the complete output of a sample "hello world" job that demonstrates this problem and we will take a look.
Re: jzos stops giving error msgs
I see what is causing the problem although this procedure has been running for several years.
I have to check the block sizes being used.
i concatenate another dataset to STDENV so that the proc can be used for all values of 'APP_HOME'.
XXSTDENV DD DSN=&&APP,DISP=(OLD,DELETE)
XX DD DSN=&SCRIPLIB(&SCRIPT),DISP=SHR
the &&APP dataset just contains a APP_HOME= parm and gets concatenated correctly:
********************************* TOP OF DATA **********************************
JVMJZBL2004N Log level has been set to: T
JVMJZBL2999T -> JzosVM()
JVMJZBL1001N JZOS batch Launcher Version: 2.3.0 2009-10-08
JVMJZBL1002N Copyright (C) IBM Corp. 2005. All rights reserved.
JVMJZBL2999T <- JzosVM()
JVMJZBL2999T -> run()
JVMJZBL1029I Region requested = 0K, Actual below/above limit = 10M / 1592M
JVMJZBL1053I OS Release R21.00 Machine 2097
JVMJZBL2999T -> adoptEnvironment()
JVMJZBL2999T -> spawnChild()
JVMJZBL1036D Spawned child shell process with PID: 579
JVMJZBL2999T <- spawnChild()
JVMJZBL2999T Writing shell script to child's stdin:
JVMJZBL2999T APP_HOME=/u/eileen/j1x/jzos/jzos_sample.jar
JVMJZBL2999T #. /etc/profile
but for some reason it throws off the rest of STDENV so that it gets ignored -
different block sizes can cause this but like I said, this has been running for years.
when i pass the data via STDENV DD *, the proc works as expected.
I have to check the block sizes being used.
i concatenate another dataset to STDENV so that the proc can be used for all values of 'APP_HOME'.
XXSTDENV DD DSN=&&APP,DISP=(OLD,DELETE)
XX DD DSN=&SCRIPLIB(&SCRIPT),DISP=SHR
the &&APP dataset just contains a APP_HOME= parm and gets concatenated correctly:
********************************* TOP OF DATA **********************************
JVMJZBL2004N Log level has been set to: T
JVMJZBL2999T -> JzosVM()
JVMJZBL1001N JZOS batch Launcher Version: 2.3.0 2009-10-08
JVMJZBL1002N Copyright (C) IBM Corp. 2005. All rights reserved.
JVMJZBL2999T <- JzosVM()
JVMJZBL2999T -> run()
JVMJZBL1029I Region requested = 0K, Actual below/above limit = 10M / 1592M
JVMJZBL1053I OS Release R21.00 Machine 2097
JVMJZBL2999T -> adoptEnvironment()
JVMJZBL2999T -> spawnChild()
JVMJZBL1036D Spawned child shell process with PID: 579
JVMJZBL2999T <- spawnChild()
JVMJZBL2999T Writing shell script to child's stdin:
JVMJZBL2999T APP_HOME=/u/eileen/j1x/jzos/jzos_sample.jar
JVMJZBL2999T #. /etc/profile
but for some reason it throws off the rest of STDENV so that it gets ignored -
different block sizes can cause this but like I said, this has been running for years.
when i pass the data via STDENV DD *, the proc works as expected.
Re: jzos stops giving error msgs
JZOS reads from DD:STDENV using the C-Library. I'm not sure why it wouldn't read your concatenation. You probably will need to open a PMR with IBM to pursue.