Re: SA78 abends
Hello Brian, $PXEQ and then $SXEQ... this will tag all active WLM initiators and shut each one down when it's batch has completed. The $SXEQ will allow new work to start in new WLM inits. Seems like I said that earlier this weekdeja vu? Regards Bruce Hewson -- 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
Re: SA78 abends - resolved
Thanks all. The $PXEQ and $SXEQ suggestions have sorted the problem of the residual state data in the WLM initiators. The problem seems to have been caused because job(s) that abended were using PDSEs. This uses DIV apparently, and some control blocks got left behind in the initiator I guess. I'm also guessing that these control blocks have some kind of ACEE/SAF association, as that would explain why I could successfully run a job under the same userid as the one that abended, but unsuccssfully run one under a different userid. Unfortunately I could not avail myself of IBM support on this problem because it's an OS/390 2.9 system (due for upgrade to z/OS 1.4 in July). Thanks again all for the help and suggestions. Brian - This e-mail message is for the sole use of the intended recipient(s)and may contain confidential and privileged information of Transaction NetworkServices. Any unauthorized review, use, disclosure or distribution isprohibited. If you are not the intended recipient, please contact thesender by reply e-mail and destroy all copies of the original message. -- 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
Re: SA78 abends
-Original Message- From: IBM Mainframe Discussion List On Behalf Of Perryman, Brian [ snip ] Sorry for not responding sooner, I'd left the office (now 10:30pm here in The 51st State). You're in Iraq?? -jc- -- 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
Re: SA78 abends
SNIP From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] Behalf Of Chase, John You're in Iraq?? UNSNIP With Blair in charge, it feels like it - similar infrastructure, but without the sunshine.. This e-mail message is for the sole use of the intended recipient(s)and may contain confidential and privileged information of Transaction NetworkServices. Any unauthorized review, use, disclosure or distribution isprohibited. If you are not the intended recipient, please contact thesender by reply e-mail and destroy all copies of the original message. -- 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
Re: SA78 abends
Thanks John, That may be in the right area, we did indeed have a lot of jobs die overnight with tape timeouts and IOS errors/abends, and also some operator cancels. However we're WLM goal mode and don't use initiators as such. I've been watching the BPXAS tasks and they have cleared down to nothing on a couple of occasions and been re-started.. I think.. Brian - This e-mail message is for the sole use of the intended recipient(s)and may contain confidential and privileged information of Transaction NetworkServices. Any unauthorized review, use, disclosure or distribution isprohibited. If you are not the intended recipient, please contact thesender by reply e-mail and destroy all copies of the original message. -- 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
Re: SA78 abends
This is weird. We have WLM-managed initiators on and all JES initiators are normally drained. WLM service classes are assigned according to the userid of the submitted batch job. If I submit a batch job with a development userid, which gets a development service class and a WLM-managed initiator, the job suffers the SA78-1C. If I submit the job with a production userid, which gets a production service class and a WLM-managed initiator, the job completes without incident. If I start a JES initiator, set MODE=JES for its jobclass, and run the job in that jobclass, it completes without incident - even with a development userid. I think I feel an IPL coming on, assuming we make it to Sunday without falling over.. Unless anyone knows any fancy WLM commands that might flush out whatever may be causing this..?? Brian - This e-mail message is for the sole use of the intended recipient(s)and may contain confidential and privileged information of Transaction NetworkServices. Any unauthorized review, use, disclosure or distribution isprohibited. If you are not the intended recipient, please contact thesender by reply e-mail and destroy all copies of the original message. -- 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
Re: SA78 abends
On Wed, 1 Jun 2005 16:57:07 +0100, Perryman, Brian [EMAIL PROTECTED] wrote: This is weird. We have WLM-managed initiators on and all JES initiators are normally drained. WLM service classes are assigned according to the userid of the submitted batch job. If I submit a batch job with a development userid, which gets a development service class and a WLM-managed initiator, the job suffers the SA78-1C. If I submit the job with a production userid, which gets a production service class and a WLM-managed initiator, the job completes without incident. If I start a JES initiator, set MODE=JES for its jobclass, and run the job in that jobclass, it completes without incident - even with a development userid. I think I feel an IPL coming on, assuming we make it to Sunday without falling over.. Unless anyone knows any fancy WLM commands that might flush out whatever may be causing this..?? $PXEQ followed by $SXEQ will drain them as they free up, but I don't know that it will help at all. Is there any software on your system that only gets its hooks involved for development jobs? Have you tried setting a slip to capture a dump to see if it provides any clues? Have you opened a PMR with IBM support? You might want to consider doing so unless you like unscheduled IPLs. :-) IPLing doesn't guarantee the problem will go away (or reappear) if you don't know what caused it. Regards, Mark -- Mark Zelden Sr. Software and Systems Architect mailto: [EMAIL PROTECTED] Systems Programming expert at http://Search390.com/ateExperts/ Mark's MVS Utilities: http://home.flash.net/~mzelden/mvsutil.html -- 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
Re: SA78 abends
IBM Mainframe Discussion List IBM-MAIN@BAMA.UA.EDU wrote on 06/01/2005 11:57:07 AM: We have WLM-managed initiators on and all JES initiators are normally drained. WLM service classes are assigned according to the userid of the submitted batch job. If I submit a batch job with a development userid, which gets a development service class and a WLM-managed initiator, the job suffers the SA78-1C. If I submit the job with a production userid, which gets a production service class and a WLM-managed initiator, the job completes without incident. If I start a JES initiator, set MODE=JES for its jobclass, and run the job in that jobclass, it completes without incident - even with a development userid. I think I feel an IPL coming on, assuming we make it to Sunday without falling over.. Unless anyone knows any fancy WLM commands that might flush out whatever may be causing this..?? I would start with SLIP SET,A=SVCD,ML=1,C=A78,RE=1C,ID=DA78,END Does the job in question use DIV? If not, maybe some DIV infrastucture somehow was not cleaned up by a job which ran in that initiator. After getting a dump, try stopping the affected initiator(s) via P INIT,A= Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY -- 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
Re: SA78 abends
If the virtual tape product Brian originally mentioned is CA:VTAPE, it is a *heavy* user of DIV. Bob Jim Mulder wrote: I would start with SLIP SET,A=SVCD,ML=1,C=A78,RE=1C,ID=DA78,END Does the job in question use DIV? If not, maybe some DIV infrastucture somehow was not cleaned up by a job which ran in that initiator. After getting a dump, try stopping the affected initiator(s) via P INIT,A= -- 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
Re: SA78 abends
SNIP__ * through termination.* * RECOMMENDATION: * Repeated ABENDA78-RC1C abends in an initiator address space. A DIV map is left outstanding which causes any subsequent subpool freemains that encompass the mapped storage to fail. This is a FINREVERSAL of APAR OW24818. --SNIP--- FINREVERSAL Any Clues? Ed -- 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
Re: SA78 abends
IBM Mainframe Discussion List IBM-MAIN@BAMA.UA.EDU wrote on 06/01/2005 04:44:38 PM: SNIP__ * through termination.* * RECOMMENDATION: * Repeated ABENDA78-RC1C abends in an initiator address space. A DIV map is left outstanding which causes any subsequent subpool freemains that encompass the mapped storage to fail. This is a FINREVERSAL of APAR OW24818. --SNIP--- FINREVERSAL Any Clues? OW24818 was closed as FIN. Subsequently, it was decided that PTFs should be created to address the problem, and a new APAR was opened for that purpose. Apparently, that process is called a FINREVERSAL - reversing the FIN closing of an APAR. Jim Mulder z/OS System Test IBM Corp. Poughkeepsie, NY -- 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
Re: SA78 abends
Ed - I believe that FINREVERSAL is used when an APAR closes FIN but then one or most customers requests that it be fixed in the current release and IBM agrees to do that. IBM documents this by opening a new APAR and marking it as a FINREVERSAL of the old APAR. Larre Shiller US Social Security Administration [EMAIL PROTECTED] V/M: (410) 965-2209 FAX: (410) 966-1936 www.ssa.gov The contents of this message are mine personally and do not necessarily reflect any official position of the US Government or the US Social Security Administration. -- 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