Re: reconcile problems

APAR IC16584 addresses the problem that you're discussing.  Please don't delete
the premigration data base.  It isn't corrupt.  The problem occurs because we
try to delete an entry that's already been deleted.  This is not an error
condition necessarily.  There are times that these entries will be deleted by
other commands while dsmreconcile is running.  IC16584 (due in V2 PTF 8, no
date yet) will fix this by logging this error and continuing on with the
dsmreconcile.

Matt Anglin
ADSM Development





ADSM-L AT VM.MARIST DOT EDU on 10-15-97 01:25:33 PM
Please respond to ADSM-L AT VM.MARIST DOT EDU @ internet
To: ADSM-L AT VM.MARIST DOT EDU @ internet
cc:
Subject: Re: reconcile problems


On Wed, 15 Oct 1997, Tor-Erik Larsen wrote:

Hi Tor-Erik,

I has the same problem for a quite long time, BTW the migration run ok,
but the dsmreconcile created the same error messages.


> I have problems with dsmreconcile.
> My reconcile process stops after updateing the candidate list. I Get
> following message on the command line:
>
>
>
> Reconciling '/data/data01' file system:
>   Reading the current premigration database...
>          Read 0 entries
>   Querying the ADSM server for a list of migrated files...
>          Received 233862 entries
>   Querying the ADSM server for a list of backed up files...
> ANS9082W dsmreconcile: error encountered while reconciling file system
> /data/data01.
>
>
> In the dsmerror log I find these messages:
>
> 10/08/1997 07:59:09  pmtIterator:LoadInCorePmigList, got a NULL from
> dbm_fetch!
> 10/08/1997 07:59:09  pmtDelete: Cannot delete record from premig database:
> /data/data01/.SpaceMan/premigrdb
> , No such file or directory
> 10/08/1997 08:36:50  DoReconcile: InitReconcile failed /data/data01, rc: 102
> 10/08/1997 08:40:47  ANS9082W dsmreconcile: error encountered while
> reconciling file system /data/data01.
>

IBM suggested this to a collegue of me:

>Ingo Grassmann wrote:
>>  18324  -  Looks like the premigration databases are corrupted.  ( I
don't know
>>  how they got that way.)  My suggestion for is for them to do the
following:
>>      o   Make sure they have the latest level of code (2.1.0.6G).
>>      o   Remove premigrdb.dir and premigrdb.pag files in .SpaceMan
directories.
>>  Then run the script "fixfsm" which will be found in /usr/lpp/adsm/bin.
>>  "fixfsm" will recreate the 2 files.  Then run reconcile.

According to this, my collegue did this on his HSM-client, and I also did
it to my client, and it worked.

I added one other precaution:

        o Make sure you copied

                cp -p premigrdb.dir premigrdb.dir.BAK
                cp -p premigrdb.pag premigrdb.pag.BAK

        o then follow the steps described before

After that the reconcile went through without any error. I asked IBM to
investigate, why the db files got corrupted, but they didn't find any
answer. One strange thing was, the corrupted premigrdb files had been VERY
large, >50MB, afer the fixed recreated files, the sizes had been normal,
about 2-10KB. But this might have been caused by the effect that this
error occured for about 2 month (I didn't call IBM, because it didn't harm
too much, everything worked as expected).


        Best regards

        Andreas

--
+----------------------------------------------------------------------+----------------------------------------------------------------------+
|       Andreas Landhaeusser                                           |
|       Deutsches Zentrum fuer Luft- und Raumfahrt e.V. (DLR)          |
|       Bunsenstr. 10                                                  |
|       D-37073 Goettingen                                             |
|       Germany                                                        |
|       Phone: +49 551 7092379  EMail: Andreas.Landhaeusser AT DLR DOT De     |
+----------------------------------------------------------------------+