Script for automate backup

[dhammika]

Hi,
I need script to backup 21 mount points. Here is my setup.
my tape library has 4 drives.
I have 21 mount points to backup.
one mount point's backup should send to one drive, that means at a time 4 mount points backup should go to 4drives separately. Then if one drive is empty(means, if one mount point's backup finished) next mount point should start to that drive. like wise all 21 mount points should backed up.

Expertise, please help me on the matter.

Thank you.

 

[moon-buddy]

Don't you have any disk pool?

The way you want the backup to run will take a long time to finish.

 

[dhammika]

Thank you for your reply...
No...there is no disk pool. The requirement is to backup to tape (This is actually flashcopy backup).
You are absolutely right, this will take lone time. Previously there was a 2 drives, but we add another 2 drives to reduce backup time. Now we have 4 drives and backing up manually. Need to automate what i am doing manually.

If there is a better way to do this please let me know.

Thank you..

 

[moon-buddy]

Adding disk pool would be nice and setting up device class as file.

The data goes to the devclass=file and later moved to tape. Faster and easy to administer.

 

[dhammika]

The problem is providing storage for disk pool, because total size of this mount points is high.
As I told you, this is FlashCopy backup and this 21 mount points mounted to the TSM server and need to distribute to drives. We already use Resource Utilization option. Then backup distributes to all 4 drives. But it is accessing one mount point at a time and taking 2 1/2 hours to finish.

Now we dont use Resource Utilization option and initiating backup manually,

dsmc archive /01 -subdir=yes
dsmc archive /02 -subdir=yes
dsmc archive /03 -subdir=yes
dsmc archive /04 -subdir=yes
.
.
dsmc archive /21 -subdir=yes

please help me to automate this task.

 

[TMan]

If I understand you correctly, you have a client with 21 different file systems which you'd like to backup efficiently to 4 tape drives in your TSM system. Perhaps this will help:

On the TSM server, do this:
upd node client_name maxnummp=4
This will let the node use 4 mount points during backup. In this case, each mount point is a tape drive.

On the client, you will need to add this line to the dsm.opt file:
RESOURCEUTILIZATION 4
This tells the BA client it is allowed to use 4 streams of data to the TSM server.

Now, if you run a straightforward incremental of the client, it will backup everything in the DOMAIN statement to four streams, going to 4 tape drives, as efficiently as it can, without you having to juggle them yourself.

Good luck!
TMan

 

[chad_small]

My understanding is that if you set resourceutilization to 4 you'll actually use 3 drive not 4. Check out the following IBM documentation:

http://publib.boulder.ibm.com/infoc...erf.doc/r_opt_client_resourceutilization.html

It appears that only when doing restores are RESOURCEUTILIZATION and MAXSESSIONS allowed a 1:1 ratio.

Also be aware that IBM recommend using filespace collocation when using multiple backup sessions on flat file backups.