ending a savsys if tape is bad
our tape is not advancing in the backup operation - how to end the job? so we can restart the backup with a new tape?
Software/Hardware used:
as/400 r5v4, ultrium tape, 3490 tape drive
Software/Hardware used:
as/400 r5v4, ultrium tape, 3490 tape drive
ASKED:
September 19, 2009 1:02 AM
UPDATED:
September 25, 2009 6:49 PM
Answer Wiki:
You can cancel a job from the WRKACTJOB screen.
If you are running interactively,
Hold down SHIFT, hit ESC
hit enter
Option 2 will end your job
----------------------------------
Most likely it's a submitted job running under QBATCH.
First you want to check the queue to see if there is anything in the queue after it that you may want to delete.
WRKJOBQ JOBQ(QBATCH)
Use 4=End or 3=Hold to remove or hold any entries you don’t want to run after you end the current backup job.
Now you can go to the QBATCH subsystem and end the current backup job.
WRKACTJOB SBS(QBATCH)
Place a 4=End beside the backup job and press F4
Change “How to end” “*CNTRLD” to “*IMMED” and press <Enter>.
Tip: You can end multiple job simultaneously by placing a 4=End beside each of the jobs you want to end then placing “OPTION(*IMMED)” on the command line and pressing <Enter>.
Voodoo
----------------------------------
To see all answers submitted to the Answer Wiki: View Answer History.
From the console press
Shift + Esc
you will get a system request line below the command line
option 2 will end your backup.
If the tape is not advancing during a SAVSYS that is because it is collecting the information to save to the tape. You need to wait for it to complete getting the information it needs and then it will advance.
You will not see the tape advance on the console screen like you do saving libraries, DLO folders or directories.
If the tape is indeed bad when the system tries to write to the tape after it gets all the data, it will return an error message and you will get a message waiting on the bottom of the screen. To see the message press Shift and ESC at the same time and place a 6 on the broken blue line at the bottom of the screen to see the messages. This will show you the error like it has a bad tape or is waiting for a tape, etc.
Hope this helps.
Lovemyi
Agree with Lovemyi.
A tape error will register a failure and kick back a message.
If you’re not seeing that then the backup is still collecting data.
If you’re running an OPT 21 save make sure you’re not selecting
Print system information . . . . Y
that will add a lot of time to your backups.
I’d advocate some more patience; if there are I/O’s assiciated with the job you can be pretty sure that it is accumulating data even if you don’t see the tape physically advancing. A lot of the time that save jobs spend is actually spent gathering lists of things to save, where they are, and so on.
Additionally, and just as anecdotal evidence only, I have had mixed results from cancelling a save job. Sometimes bad things happen; jobs hang forever (or seemingly so), device descriptions get damaged, I cancel a save job only as an absolute last resort.
If the current time for the SAVSYS is 2 or 3 times longer than normal and you have not installed any group or single PTFs, then you probably have an issue. In the past, i’ve had the operator eject the tape and hopefully invoke an error. On perhaps 3 occasions it did (unable to write to tape i believe?) and the operator just retried and all was well. One instance though, after waiting another 20 mintues, he called IBM and the rep suggested an IPL. The SAVSYS ran fine thereafter. Since all 4 eventually completed normally (though it ran 4 times longer than the norm), IBM never gave a satisfactiory explanation and repeated follow up calls produced nothing. No surprise there.
There have been a lot of Save/Restore PTFs out for v5r4 lately. I would make sure I was current on PTFs. Also, take the recommendations about waiting a reasonable time before ending the job to heart. iSeries saves mark each object with the last save iinformation as the save occurs. When you end a save job it has to reverse all of that information back off the objects as the iSeries is very serious about data integrity. It is best to get the job to error and give it a retry with a new tape otherwise it can take as long to end as it did to get to the point of failure.