Yes, that would explain your problem. I would do archiving even on your
test system in order to avoid problems like this. You could specify a
retention period of just a day or two in order to avoid tying up alot of
space.
- Bruce Giordano
"Jose, Prince" <[EMAIL PROTECTED]>
To:
[EMAIL PROTECTED]
Sent by: MQSeries List cc:
<[EMAIL PROTECTED]> Subject: Re: Qmgr abend - 5C6
Wednesday June 11, 2003 05:57 PM
Please respond to MQSeries List
Bruce, It makes sense now.
This is our test system and we do not use archiving or dual logging.
The programmer was doing load testing with persistent messages and the
batch
job abended.
(also, commit is not used in the program!)
Since circular logging is used in mainframe, MQ might have overwrote the
log datasets.
And when the job abended, MQ tried to backout the changes, but could not
find the original log record.
And abended...
Does it make sense?
Thanks, Prince
-----Original Message-----
From: Bruce Giordano [mailto:[EMAIL PROTECTED]
Sent: Wednesday, June 11, 2003 3:38 PM
To: [EMAIL PROTECTED]
Subject: Re: Qmgr abend - 5C6
Looks like the batch job abended and MQSeries is trying to back out its
activity. Did a log archive occur while this batch job was putting its
messages? If so, is it possible that the archive file was deleted? If not
I think I'd open a problem with IBM. If one logdataset is truely corrupt
and you are using dual logging, you should be able to copy the good log
dataset into the bad.
- Bruce Giordano
"Jose, Prince" <[EMAIL PROTECTED]>
To:
[EMAIL PROTECTED]
Sent by: MQSeries List cc:
<[EMAIL PROTECTED]> Subject: Qmgr
abend - 5C6
Wednesday June 11, 2003 03:00 PM
Please respond to MQSeries List
Hello!
Our queue manager on Mainframe (5.2) abended yesterday with an SVC dump. (
5C6)
After going through the logs and looking at error message description,
looked to me that the log dataset got corrupted.
The user id appeared on the dump(ABC12) was running a batch job putting
persistant messages on to the queue.
(1000 msgs 100k in size each)
Can anyone tell me what can cause the logdatset to get corrupted?
Thanks, Prince
Here are the relevent MQ error messages as it appeared in the syslog.
CSQ3201E *MQSB ABNORMAL EOT IN PROGRESS FOR 830
USER=ABC12 CONNECTION-ID=ABC12RN2
THREAD-XREF=000000000000000000000000
CSQJ113E *MQSB RBA 00001C637000 NOT IN ANY ACTIVE OR 944
ARCHIVE LOG DATA SET, CONNECTION-ID=MQSB THREAD-XREF=
IEA794I SVC DUMP HAS CAPTURED: 980
DUMPID=005 REQUESTED BY JOB (MQSBMSTR)
DUMP TITLE=MQSB,ABN=5C6-00D1032A,U=ABC12 ,C=F1000.520.RLMC-CS
QJLGR ,M=CSQJRE01,LOC=CSQJL002.CSQJR003+0582
CSQV086E *MQSB MQSERIES ABNORMAL TERMINATION REASON=00D94001
+CSQX411I *MQSB CSQXREPO Repository manager stopped
Prince Jose
Project Support Resources - (Middleware Products)
Phone : 313 225 8156
(See attached file: C.htm)
(See attached file: C.htm)
Title: RE: Qmgr abend - 5C6
Bruce, It makes sense now.
This is our test system and we do not use archiving or dual logging.
The programmer was doing load testing with persistent messages and the batch job abended.
(also, commit is not used in the program!)
Since circular logging is used in mainframe, MQ might have overwrote the log datasets.
And when the job abended, MQ tried to backout the changes, but could not find the original log record.
And abended...
Does it make sense?
Thanks, Prince
-----Original Message-----
From: Bruce Giordano [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, June 11, 2003 3:38 PM
To: [EMAIL PROTECTED]
Subject: Re: Qmgr abend - 5C6
Looks like the batch job abended and MQSeries is trying to back out its
activity. Did a log archive occur while this batch job was putting its
messages? If so, is it possible that the archive file was deleted? If not
I think I'd open a problem with IBM. If one logdataset is truely corrupt
and you are using dual logging, you should be able to copy the good log
dataset into the bad.
- Bruce Giordano
"Jose, Prince" <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Sent by: MQSeries List cc:
<[EMAIL PROTECTED]> Subject: Qmgr abend - 5C6
Wednesday June 11, 2003 03:00 PM
Please respond to MQSeries List
Hello!
Our queue manager on Mainframe (5.2) abended yesterday with an SVC dump. (
5C6)
After going through the logs and looking at error message description,
looked to me that the log dataset got corrupted.
The user id appeared on the dump(ABC12) was running a batch job putting
persistant messages on to the queue.
(1000 msgs 100k in size each)
Can anyone tell me what can cause the logdatset to get corrupted?
Thanks, Prince
Here are the relevent MQ error messages as it appeared in the syslog.
CSQ3201E *MQSB ABNORMAL EOT IN PROGRESS FOR 830
USER=ABC12 CONNECTION-ID=ABC12RN2
THREAD-XREF=000000000000000000000000
CSQJ113E *MQSB RBA 00001C637000 NOT IN ANY ACTIVE OR 944
ARCHIVE LOG DATA SET, CONNECTION-ID=MQSB THREAD-XREF=
IEA794I SVC DUMP HAS CAPTURED: 980
DUMPID=005 REQUESTED BY JOB (MQSBMSTR)
DUMP TITLE=MQSB,ABN=5C6-00D1032A,U=ABC12 ,C=F1000.520.RLMC-CS
QJLGR ,M=CSQJRE01,LOC=CSQJL002.CSQJR003+0582
CSQV086E *MQSB MQSERIES ABNORMAL TERMINATION REASON=00D94001
+CSQX411I *MQSB CSQXREPO Repository manager stopped
Prince Jose
Project Support Resources - (Middleware Products)
Phone : 313 225 8156
(See attached file: C.htm)
