ADSM-L

Re: TSM migration problem & recovery log question???

2003-01-14 11:07:12
Subject: Re: TSM migration problem & recovery log question???
From: Joni Moyer <joni.moyer AT HIGHMARK DOT COM>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Tue, 14 Jan 2003 11:05:25 -0500
I really don't know what is wrong with the recovery log.  There are no
clients that have been connected to the server for more than 4 minutes.  I
have a backup of a primary tape pool to a copy tape pool running 1 process.
Other than that everything is silent.  I can't seem to get the log to clear
to 0% and soon I'll have to add more space which will take it to 6 GB for
the recovery log.  I am still baffled about this.

We have TSM on the mainframe and it is attached to STK silos with 3590
magstar tapes.  As far as I know communication wasn't lost because other
migrations/backups etc. were running.  I was just having problems with 1
out of 3 of my disk storage pools migrating to tape.  I looked through the
logs trying to find out what tape it might have been trying to migrate to,
but it never asked for a certain tape and I never received a mount request
either. I know this all seems very bizarre.  I really don't know how to
justify what happened.  Thanks!!!

Joni Moyer
Systems Programmer
joni.moyer AT highmark DOT com
(717)975-8338



                    "Cook, Dwight
                    E"                   To:     ADSM-L AT VM.MARIST DOT EDU
                    <DWIGHT.E.COOK       cc:
                    @SAIC.COM>           Subject:     Re: TSM migration problem 
& recovery log question???
                    Sent by:
                    "ADSM: Dist
                    Stor Manager"
                    <ADSM-L AT VM DOT MAR
                    IST.EDU>


                    01/14/2003
                    09:20 AM
                    Please respond
                    to "ADSM: Dist
                    Stor Manager"






Because of internal locks that aren't being readily resolved.
Look for a client (or clients) that have been connected for a long time
(say
4-6+ hours)
Also look for other things like expiration, reclamation, migration, etc...
processes that might be running also.
Might only been client sessions though...
Anyway, generally (I've discovered) the log won't clear until those long
running clients either finish normally OR I cancel them.
We have had to increase some of our log sizes to 6 GB (6144 MB) in order to
allow things to run to completion and not have our log fill up and cause
sessions to die off due to log space being exhausted.

Now, media inaccessible, how is your library attached ?????
I've noticed (in the past) that if the library manager can't talk to TSM
(or
tsm can't talk to the library manager) AND tsm requires a tape mount (which
fails because of lack of communication) even though I restored
communications to the library, I would still have to bounce TSM to get it
to
start talking to the library again...
Is there something that is causing a loss of communications with your
library prior to this "media not available" situation ?
When things fail due to media not available does it report a specific
volser
?
When things are bounced and start working again, does the problem volser
(if
identified) become accessible ?

Dwight E. Cook
Software Application Engineer III
Science Applications International Corporation
509 S. Boston Ave.  Suite 220
Tulsa, Oklahoma 74103-4606
Office (918) 732-7109



-----Original Message-----
From: Joni Moyer [mailto:joni.moyer AT HIGHMARK DOT COM]
Sent: Tuesday, January 14, 2003 7:23 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: TSM migration problem & recovery log question???


The recovery log on TSM is 95% full.  I do a full DB backup everyday, which
I thought would reset the log to 0%.  It is in normal mode, not
rollforward.  Why would it be so high when I just did a full backup last
night that ended at 9 PM.  The log right now is 4.6 GB with 4.2 GB
utilized.  Any suggestions?

We have TSM 4.1.3 on OS/390.  My other problem is that our disk storage
pool filled to 100%, getting message anr1021w: media inaccessible.  When
trying to force migration (which should've occurred automatically) the
processes still failed with the media inaccessible messages.  We were not
in tape drive allocation and plenty of scratch volumes were available.  No
messages in the logs even pointed to a specific tape that was being called
in.  Bouncing the tsm server triggered one of four possible migration
processes.  Around 8 hours after bouncing the server 4 migration processes
finally did kick off after hitting the set threshold.  Any ideas or insight
as to what the problem could be???

Thanks for any help,

Joni Moyer
Systems Programmer
joni.moyer AT highmark DOT com
(717)975-8338

<Prev in Thread] Current Thread [Next in Thread>