Backup primary pool to copypool

mkraker

ADSM.ORG Member
Joined
Feb 9, 2011
Messages
20
Reaction score
1
Points
0
PREDATAR Control23

Goodday all,

I have a primary pool (devclass DISK) en i want to make a backup to copypool (devclass FILE s, sequential access).

Everytime the backup is started a 50GB sequential file is created.

When i backup the primary to the copy pool , after the 1st volume of 50G reaches 0.3 utilitisation , it becomes full , and started the backup to the 2nd 50GB volume.

I have the copypool set to COLL=NO
MAXCAP=50G
MAXSCRATCH=200

This way i am never ever able to backup my primary pool.

Am i missing here something or am i doing something wrong?

Thx for any reply.

Kind regards,

Michel.
 
PREDATAR Control23

Goodday all,

I have a primary pool (devclass DISK) en i want to make a backup to copypool (devclass FILE s, sequential access).

Everytime the backup is started a 50GB sequential file is created.

When i backup the primary to the copy pool , after the 1st volume of 50G reaches 0.3 utilitisation , it becomes full , and started the backup to the 2nd 50GB volume.

I have the copypool set to COLL=NO
MAXCAP=50G
MAXSCRATCH=200

This way i am never ever able to backup my primary pool.

Am i missing here something or am i doing something wrong?

Thx for any reply.

Kind regards,

Michel.


50 GB is too small - try increasing to 100 GB or more. I have a similar setup at 100 GB per volume and all goes well.
 
PREDATAR Control23

same issue
And what exactly is the issue? I understand you cannot finish the backup stgpool, but that's rather vague. What errors does the backup stgpool finish with? That would be the first piece of info needed to start troubleshooting.
 
PREDATAR Control23

And what exactly is the issue? I understand you cannot finish the backup stgpool, but that's rather vague. What errors does the backup stgpool finish with? That would be the first piece of info needed to start troubleshooting.

The problem is that a 50GB or 100GB volume gets the status full after it gets approx 250MB utilised. So i created 14 x 100GB volume , but with this , i am able to backup 14 x 250MB because after a while all 14 volumes reporting FULL with q vol.
No error in the actlog on this , except offcourse after the 14th volume the message is that the copypool is out of space.

The primary pool is a diskpool . Before i started the backup to the copypool i deleted one big Filespace from the primary pool.
I do not know if this causes this , but maybe it is.
 
PREDATAR Control23

The primary pool is a diskpool . Before i started the backup to the copypool i deleted one big Filespace from the primary pool.
I do not know if this causes this , but maybe it is.
Actually, the opposite, it should help because you'll have less data to backup.

What's your server version?
 
PREDATAR Control23

hi Marclant. I opened a SR@IBM for this issue. Turned out , i need to use the DATAFORMAT=NONBLOCK while creating the copypool. Now it is running fine. The answer from IBM:

First i need you to know something very important about the File Storage Pool, in this environment the minimum blocksize used for Tivoli Storage Manager Server FILE device types is 256K. When aggregation is not possible, each object, no matter how small will occupy/require a full block of space in the FILE storage pool.

For example: a 700 byte directory backed up as a single object will
require a full 256K block of space in the device type FILE volume. All
the space in the 256K block not occupied by the 700 bytes of the
directory data is lost (physically not available for use).

So, i think that you are backing up a small objects which as per the
above reason is filling the Volume rapidly without Physically reaching
the maximum size.

The solution :)

Yes there is a solution and it is a changing the File pool from Block
to Non Block, you will need to redefine the storage pool as the update
command cannot change the "dataformat" parameter.
 
PREDATAR Control23

hi Marclant. I opened a SR@IBM for this issue. Turned out , i need to use the DATAFORMAT=NONBLOCK while creating the copypool. Now it is running fine. The answer from IBM:

First i need you to know something very important about the File Storage Pool, in this environment the minimum blocksize used for Tivoli Storage Manager Server FILE device types is 256K. When aggregation is not possible, each object, no matter how small will occupy/require a full block of space in the FILE storage pool.

For example: a 700 byte directory backed up as a single object will
require a full 256K block of space in the device type FILE volume. All
the space in the 256K block not occupied by the 700 bytes of the
directory data is lost (physically not available for use).

So, i think that you are backing up a small objects which as per the
above reason is filling the Volume rapidly without Physically reaching
the maximum size.

The solution :)

Yes there is a solution and it is a changing the File pool from Block
to Non Block, you will need to redefine the storage pool as the update
command cannot change the "dataformat" parameter.

This should not be an issue as the default setting of NATIVE would not constrain data allocation to a certain block size.

I leave the default to NATIVE and all works well.
 
Top