Some of this will depend upon the size of your SAP filesystems and on your
MULTIPLEXING value. TSM sees the sum of <multiplexing> <filesystem sizes> as
the actual 'file size' for space calculation. In my case, 1.6 GB files (all
of them the same size) and 6-way multiplexing, for a 9.6 GB data file to
TSM.
When you reach your migration percentage high-point for the storage pool,
TSM will kick off the number of migration processes you've specified. When
the free space in the pool drops below the data file size the session will
switch to your next pool -- the tape pool. If you still have drives
available, you'll get a mount and continue. If all drives are in use, the
session will wait for a drive to free up *even if the migration process
frees up disk space* for the TSM file it is working on.
You may want to try both with and without the disk pool. I'm not sure an 80
GB disk pool is worth the effort for a 600 GB database. We run 4 sessions
direct to LTO tape drives and back up around 600 GB in about 3 hours (two
sessions each on two gigabit ethernet networks).
It all boils down to an exercise in finding and fixing bottlenecks until you
reach a performance level you can live with or hit a bottleneck management
doesn't want to spend the money to fix. These include disk read rate from
the SAP DB (disk drives, disk drive interface, cpu); network from DB server
to TSM server (hubs, switches, interface speed, protocol, cpu, interface
cards); TSM server disk performance (read AND write); TSM tape interface;
tape drives. And I left a few out -- these are just the areas I've played
with in the last five years.
Hope this helps a bit --
Tom Kauffman
NIBCO, Inc
> -----Original Message-----
> From: Van Ruler, Ruud R SITI-ITDSES64 [mailto:ruud.vanruler AT SHELL DOT COM]
> Sent: Friday, March 29, 2002 8:44 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: SAP R3 and (MAX_)SESSIONS
>
>
> Hi
>
> environment:
> one TSM server with disk pool stg_sapdisk_1 (Estimated
> Capacity (MB) 80.000
> i.e. 80GB).
> next storage pool defined for stg_sapdisk_1 is tape (6 drives are
> available).
>
> every two days a tdp for r/3 backup runs (sap environment is 600Gb).
> I have set up R3 client for 6 parallel sessions to disk storage pool
> stg_sapdisk_1 (consists of 20 disk volumes).
>
> at some point (as you can imagine) the disk pool is filled up
> and migration
> will be performed automatically.
> what will happen next ?
> are current SAP backup sessions redirected straight to tape ??
> will the current sessions wait until there is space again on
> the disk pool
> ??
> there is only data from this one R/3 client on the disk pool
> .... so there
> will be only one migration process started ??
>
> suppose this stg_sapdisk_1 consists of 3 disk volumes. Can I
> assume that
> then 3 sessions go to disk pool and the other remaining 3 go
> to the next
> storage pool i.e. tape ??
>
>
>
>
>
> Ruud van Ruler, Shell Information Technology International
> B.V. - DSES-64
> ITDSES-6 Technical Information and Links page:
http://pat0006/shell.htm
Room 1A/G03
Dokter van Zeelandstraat 1, 2285 BD Leidschendam NL
Tel : +31 (0)70 - 3034644, Fax 4011, Mobile +31 (0)6-55127646
Email Internet: ruud.vanruler AT shell DOT com
rvanruler AT hotmail DOT com
|
