Re: TDP SQL sessions wants to write to same tape...
Goran is correct. You need to make sure that MAXNUMMP for the
NODENAME you are using for your SQL backups is set to at least
to the number of stripes you are using. Also, turning on
collocation by filespace will ensure that data for different stripes
remains on different tapes so that the restores will also be able to
read from separate tapes at the same time.
>From Chapter 3 of the User's Guide:
- The maxnummp parameter for register or update node must be at least the
maximum number of stripes to be used for backup or restore when
media such as tapes are used or if migration occurs during the backup or
operation. The default value is 1. This value is the maximum number of
points a node can have at one time. If other backups or restores may
the same time, the value of this parameter must be large enough to allow
of the needed mount points.
- The collocate parameter for define stgpool should be set to "filespace"
that individual data stripes stay on separate removable volumes. The
value is no. Not doing so may require move operations to make the data
each stripe of a restore simultaneously accessible.
There is also a nice IBM KB article (1145253) describing more details. Go
Search for this (include the double quotes in the search window):
"DP for SQL : Data Striping"
"ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU> wrote on 02/13/2004
> I've been running a backup of a SQL server db but I took too long so
> I increased the stripe-parameter to 2 so it would utilize 2 sessions
> and 2 tape drives for the backup. However when the backup starts
> both sessions wants to write to the same volume, leaving one of the
> session in a media-wait state for the whole backup and the backup
> fails after 12h.
> 12,064 BPX-T- RecvW 0 S 20.5 K 62.6 G Node TDP MSS- ROBUR22_SQL
> cp/Ip QL
> 12,066 BPX-T- MediaW 13.0 H 5.2 K 2.5 M Node TDP MSS- ROBUR22_SQL
> cp/Ip QL
> What can I do to fix this???
> Niklas Lundström