I had this problem too. I solved doing the fourth option. Maybe after you
change the MAXNUMMP to zero happens the sessions of the clients finishes
before all the backup with failure. I solved this other problem setting the
parameter IDLETIMEOUT to 60. This problem happens because if the disk
stgpool fulls first than the data migrates from disk to tape. So the clients
waits until the process of migration finishes and this can be more than 15
minutes.
Regards,
Anderson
> There are several things you can try to fix it, depending on which works
> better for you:
>
> 1) Set the high-mig value for the disk pool LOWER. Or make your disk
pool
> BIGGER.
> Either way, to make sure that when the client wants to back up, there is
> always enough room for him in the disk pool
>
> 2) Change your scheduling so that the biggest client backs up FIRST; so
> there will be plenty of room for him in the disk pool.
>
> 3) Set up an admin schedule that starts a couple of hours before the
client
> backup, and changes the high-mig value to 0. This will force migration to
> start early, so the disk pool empties out before your big client starts
its
> backup.
>
> 4) Change the MAXNUMMP (max number of mount points) for the client to 0.
> That means the client is NOT allowed to grab a tape drive for backup.
> However, if you do this and there is not enough room in the disk pool, the
> client backup will fail.
>
> ************************************************************************
> Wanda Prather
> The Johns Hopkins Applied Physics Lab
> 443-778-8769
> wanda_prather AT jhuapl DOT edu
>
> "Intelligence has much less practical application than you'd think" -
> Scott Adams/Dilbert
> ************************************************************************
>
>
>
>
>
> -----Original Message-----
> From: Gerhard Wolkerstorfer [mailto:Gerhard.Wolkerstorfer AT T-SYSTEMS DOT AT]
> Sent: Friday, December 21, 2001 7:55 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Antwort: Migration
>
>
> It could be, that the available FREE space in the diskpool at this time is
> smaller, than the size of the file your client want to send to the server
> This could be the reason, why the server directs the data directly to
> tape...
>
> -- Gerhard --
>
>
>
>
>
> Christoph.Pilgram AT BC.BOEHRINGER-INGELHEIM DOT COM (Christoph Pilgram) am
> 21.12.2001
> 11:12:08
>
> Bitte antworten an ADSM-L AT VM.MARIST DOT EDU
>
> An: ADSM-L AT VM.MARIST DOT EDU
> Kopie: (Blindkopie: Gerhard Wolkerstorfer/DEBIS/EDVG/AT)
>
> Thema: Migration
>
>
>
> Hi all,
>
> today I have a question dealing with migration :
>
> I have one disk-storage-pool that is filled up during the night by several
> clients. When it reaches the High-Mig-Pct, the Migration starts with up to
> three jobs going to tape.
> Backups havent finished at that time and suddenly the backup-jobs wants a
> tape to be mounted, and this is just that tape, that is mounted by the
> migration-job (because of collocation on client). So the backup-job has to
> wait and wait.
> But the question is, why the backup of the client, that has as
primary-pool
> only a disk-pool whis no "maximum size threshhold" transfers data direct
to
> tape (while disk pool is migrating) and whow to stop this behaviour.
>
> Thanks for help
>
> Chris
>
> Boehringer Ingelheim Pharma KG
> IT Department
> Christoph Pilgram Tel. +49 7351 54 4051
> Birkendorfer Str.65 Fax. +49 7351 83 4051
> 88397 Biberach Email :
> Christoph.Pilgram AT bc.boehringer-ingelheim DOT com
> Germany
>
|
