HardeepSingh
ADSM.ORG Member
Hi,
We're facing a lot of Capacity issues recently which led us to investigate further on why we are not deleting data as much as we are adding in.
Case 1:
One of the major concerns we are facing is due to the haphazardness of the setup.
We have multiple Data Domains setup for our environment, each serving the purpose of vaulting and replication for each other. Now, since the early days, TDP for Oracle was pointing to a copy destination directly backing to the NFS share at the DD. Lets just call it DD1.
Now, we also had VTL setup for the same TSM server, but was using the VTL of DD2.
Now DD1 and DD3 are each others replication or DR site
And DD2 and DD4 are each others replication or DR site.
So, when DD1 started getting full, we sought space from DD2 and moved the backup to VTL. Once VTL started getting full after 2-3 months, we moved the destination back to DD1 as it was able to reclaim some space by this time. This activity has been repeated over 4-5 times now. Policy domain remained the same.
With the backup retention advised as 21 days to the RMAN DBAs, they continue expiring the data. But what we've noticed is, the data doesn't completely expire during that 2-3 month window. We still see tapes holding data of random months say, May, August, October of 2013 and Feb, Apr of 2014. Doesn't make sense?
If the copy destination keeps changing, does it affect expiration of data on TSM?
Case 2:
We've changed the policy domain of a number of nodes recently, allowing them to change to a new copy destination. These nodes were causing issues with backup as the local STG pool would fill up to 100% quickly once these started backing up.
Keeping all copygroup parameters the same, we've changed the domain and copy destination of these nodes.
Will this affect expiration & restoration process of the data?
We're facing a lot of Capacity issues recently which led us to investigate further on why we are not deleting data as much as we are adding in.
Case 1:
One of the major concerns we are facing is due to the haphazardness of the setup.
We have multiple Data Domains setup for our environment, each serving the purpose of vaulting and replication for each other. Now, since the early days, TDP for Oracle was pointing to a copy destination directly backing to the NFS share at the DD. Lets just call it DD1.
Now, we also had VTL setup for the same TSM server, but was using the VTL of DD2.
Now DD1 and DD3 are each others replication or DR site
And DD2 and DD4 are each others replication or DR site.
So, when DD1 started getting full, we sought space from DD2 and moved the backup to VTL. Once VTL started getting full after 2-3 months, we moved the destination back to DD1 as it was able to reclaim some space by this time. This activity has been repeated over 4-5 times now. Policy domain remained the same.
With the backup retention advised as 21 days to the RMAN DBAs, they continue expiring the data. But what we've noticed is, the data doesn't completely expire during that 2-3 month window. We still see tapes holding data of random months say, May, August, October of 2013 and Feb, Apr of 2014. Doesn't make sense?
If the copy destination keeps changing, does it affect expiration of data on TSM?
Case 2:
We've changed the policy domain of a number of nodes recently, allowing them to change to a new copy destination. These nodes were causing issues with backup as the local STG pool would fill up to 100% quickly once these started backing up.
Keeping all copygroup parameters the same, we've changed the domain and copy destination of these nodes.
Will this affect expiration & restoration process of the data?