Message interception during iSeries backup
Our site (where I have been working just a few days) had a backup failure on the iSeries last night. An operator simply forgot to load the tape, but this resulted in the machine alerting the SYSOPR and waiting - all night - for a response. Since no one was around to answer the message, the backup didn't happen and the machine was unavailable when staff came in this morning. Without using a CL program with MONMSG commands embedded in it, how can this situation be handled? I'd love to implement the very neat looking Operational Assistant options for backup, but don't see any way for monitoring for messages that might arise. I might not be able to save the night's backup, but at least the system would be available the next day.
Software/Hardware used:
iSeries model 525
Software/Hardware used:
iSeries model 525
ASKED:
September 6, 2012 7:55 PM
UPDATED:
September 7, 2012 6:07 PM
Answer Wiki:
Last Wiki Answer Submitted:
Be the first to answer this question.
All Answer Wiki Contributors:
Be the first to answer this question.
To see all answers submitted to the Answer Wiki: View Answer History.
What would you do if you caught a message? Hundreds of different messages might show up. — Tom
Why not use a CL with a MONMSG?We had a similar issue here. Some batch jobs would halt after hours and stop the backups from running. All we did to resole was monitor for the messages and if the came up had the system send an e-mail message to the person responsible for the halt or backups so they could resolve it.
Get the message ID and change the reply list for that message. ADDRPYLE, CHGRPYLE to INZ the tape. Problem is, that will always be the answer and you will get no second chances. You could initialize a tape that should not have been initialized. But, then, it should not have been in the tape drive to be used for the backup.Nick
Problem is, that will always be the answer and you will get no second chances.
It should only affect a job that has INQMSGRPY(*SYSRPYL) in effect when the message is sent. And it will only match if the comparison data matches the message data.
Control of those allows the choice of when the system reply list makes a difference.
Tom
If you have no program to monitor the processing error from the night batch job ,I think it’s a huge risk.You can use email or send a call by the modem to the duty guy to response the alert.
For the night backup,we backup the data to a savf first,and then sbmjob to save the savf to tape, if the tape is not ready,you can reply the message in the second day.I think the way by savf is less mistake than directly save to the tape.
To minimize the problem of the tape not loaded, we run a Job Schedule Entry job late in the work day that checks if a tape is loaded. It warns if no tape is found. This is just a the critical part of the code:CHKTAP DEV(&DEV) VOL(*MOUNTED) ENDOPT(*REWIND)
MONMSG MSGID(CPF6760 CPF6772) EXEC(DO) /* NO TAPE +
IN UNIT */
SNDBRKMSG MSG(‘** i5: No tape in drive or tape not +
initialized **’) TOMSGQ(&MSGQ1 &MSGQ2 +
&MSGQ3 &MSGQ4)
MONMSG MSGID(CPF0000)
ENDDO
It has saved our bacon a few times.
Sorry . . .The code snippit posted differently than it appeared when I keyed it.
Something else you might consider is SndTweet by Kisco. We use it to send a tweet when an error message requiring a reply lands in QSYSOPR. The cool thing about SndTweet is that we can respond to the message via Twitter. Cost is very reasonable. It application runs entirely on the i5 – no separate Wintel servers required.
WoodEnginee, I like that approach. Fix it before it happens.