ADSM-L

Re: Strange Reclamation Problem

2006-09-19 12:19:02
Subject: Re: Strange Reclamation Problem
From: Andrew Carlson <naclos AT SWBELL DOT NET>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Tue, 19 Sep 2006 09:16:27 -0700
Ok, some additional information.

If i repeatedly reclaim or move/data the volume, more data gets moved.  The 
next nodes amount of data gets moved.  The disks are fine, no errors, no 
destroyed volumes, no corruption that I can see.  It just seems that TSM thinks 
it is done and ends the reclamation process.  The next one starts up and moves 
some data, and ends  like it ended in success, but still has not moved all the 
data.  Eventually, after many many tape mounts and multiple reclamation 
processes, the tape is finally reclamated, but it's just not very efficient.

I think I am going to have to open a PMR with IBM - this feels more like a 
defect than anything else - was hoping some else had seen this problem.

Andy Carlson 
---------------------------------------------------------------------------
Gamecube:$150,PSO:$50,Broadband Adapter: $35, Hunters License: $8.95/month,
The feeling of seeing the red box with the item you want in it:Priceless.

----- Original Message ----
From: Richard Sims <rbs AT BU DOT EDU>
To: ADSM-L AT VM.MARIST DOT EDU
Sent: Tuesday, September 19, 2006 10:06:39 AM
Subject: Re: [ADSM-L] Strange Reclamation Problem

Well, I don't know how you're making onsite primary disk storage
pools offsite.  There seem to remain details missing from your
conveyed configuration.  But, moving on...

A reclamation of offsite-marked tapes occurs through the utilization
of onsite data, generically speaking.  The only conventional reason
for that to fail to complete is that not all the onsite data is
available.  There's a psychology in effect where TSM administrators
expect tape volumes to go bad, and they accept that as a daily
reality and deal with that as a matter of course.  But, when the
volumes are disk, they are assumed to all be fully operational and
are almost never examined; and even when something is wrong, the
state of the disk volumes remains unexplored.  The circumstances seem
to me to suggest some issues with the disk volumes there, warranting
a long, hard look in this instance, and ongoing monitoring for the
long term.

    Richard Sims

On Sep 19, 2006, at 10:10 AM, Andrew Carlson wrote:

> Richard,
>
> I think it was obvious, but maybe it wasn't.  People don't normally
> make their onsite disk pools into offsite status.  When my copypool
> tapes get full, I mark them offsite so that reclamation will read
> the disk pool, not do reclamation tape to tape.  But, this is what
> is not working.  Reclamation starts, moves a little bit of data,
> assumably one node's worth from what I saw in move data, then ends
> with "success", not fully emptying the tape that is being reclamated.
>
> Andy Carlson
> ----------------------------------------------------------------------
> -----
> Gamecube:$150,PSO:$50,Broadband Adapter: $35, Hunters License:
> $8.95/month,
> The feeling of seeing the red box with the item you want in
> it:Priceless.
>
> ----- Original Message ----
> From: Richard Sims <rbs AT bu DOT edu>
> To: Andrew Carlson <naclos AT swbell DOT net>
> Sent: Monday, September 18, 2006 5:07:20 PM
> Subject: Re: Strange Reclamation Problem
>
> Andy -
>
> I don't think you're going to get many replies, in that the posting
> is too incomprehensible...
>  From the posting, we have no idea what you're changing to read-
> write, as your posting
> talks of only a primary disk pool and a 3584 containing 3592 tapes
> for your offsite pool.
> I read the posting four times, and still don't understand it.
>
> You may want to post a re-writing, to fully explain the apparent
> missing ingredients.
> Note that message ANR1163W is explained, for the general case, in
> ADSM QuickFacts.
>
>     Richard Sims
>
> On Sep 18, 2006, at 4:07 PM, Andrew Carlson wrote:
>
>> I am running TSM 5.3.2.3 on an AIX (5.2.5) platform.  We recently
>> finished migrating our onsite pool to disk only, and out offsite
>> pool to directly attached 3592 drives in a 3584 silo.  We started
>> getting multiple ANR1163I messages.  I started investigating this,
>> and found an odd behaviour.  If I entered a move volume command for
>> one of the volumes, it would move some data, then end, with no
>> errors.  It appears to be moving the data of one node off the cart,
>> then ending the process.  These are volumes that were previously
>> not collocated.  If I update the volume to read-write (I forgot to
>> mention I make them offsite so the TSM will read from the disk
>> pool), and do the move volume tape-to-tape, it works fine.  It
>> finishes the whole tape.  Any ideas?  I am going to open a PMR with
>> IBM, but thought I would ask here first in case I missed something
>> incredibly obvious.  Thanks.
>>
>> ANR1040I Space reclamation started for volume 110021, storage pool
>> COPY3592
>> (process number 1555).
>> ANR1040I Space reclamation started for volume 110482, storage pool
>> COPY3592
>> (process number 1555).
>> ANR1040I Space reclamation started for volume 110436, storage pool
>> COPY3592
>> (process number 1555).
>> ANR1040I Space reclamation started for volume 110268, storage pool
>> COPY3592
>> (process number 1555).
>> ANR1040I Space reclamation started for volume 111027, storage pool
>> COPY3592
>> (process number 1555).
>> ANR8468I 3592 volume 110192 dismounted from drive RMT5 (/dev/rmt5)
>> in library
>> 3584LIB.
>> ANR0409I Session 349997 ended for server TSMLIBM (AIX-RS/6000).
>> ANR1163W Offsite volume 110212 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110499 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110163 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110021 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110482 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110436 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 110268 still contains files which could not
>> be moved.
>> ANR1163W Offsite volume 111027 still contains files which could not
>> be moved.
>> ANR4932I Reclamation process 1555 ended for storage pool COPY3592.
>> ANR0986I Process 1555 for SPACE RECLAMATION running in the
>> BACKGROUND processed
>> 6335 items for a total of 1,768,794,536 bytes with a completion
>> state of
>> SUCCESS at 10:13:52.
>>
>>
>> Andy Carlson
>> ---------------------------------------------------------------------
>> -
>> -----
>> Gamecube:$150,PSO:$50,Broadband Adapter: $35, Hunters License:
>> $8.95/month,
>> The feeling of seeing the red box with the item you want in
>> it:Priceless.

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