I am familar with creating ddboost storage units, and making mtree replications out of them. BUT, what i havent done before was to take an existing ddboost storage unit that contains data, and replicate that to my secondary datadomain. Is it possible to do this? If so, is it dangerous? I mean, if its done, could the new (empty) mtree side actually replicate its "emptiness" to the existing storage unit that contains data? If mtree is not an optimal method since the main storage unit already contains data, is there a better option?
As you use NW and DDBoost, why simply not use CCR? Given that this gives you full application awareness of replicated bits...
Hi Hrvoje. The issue is that the data in this storage unit is Oracle database data written with RMAN DDBOOST, so that data doesnt come through Networker.
so you have a ddboost container that is being written to by Oracle RMAN Boost ? I told my DBAs to stop RMAN jobs for a second, i create mtree replication context, initialized it ..and told DBA they could resume their jobs. Mtree got automatically create on target DD. On the day of cut-over, DBAs again stopped RMAN jobs, i waited for replication to get to "synchronized" state and then broke replication context. On target DD i assigned DDBoost user to ddboost container, at that point it became visible as a storage-unit. DBA re-ran their DDBoost registration script with updated DD host name and that's it.
Thats kind of a fear that i have, using mtree for this. Im concerned that since mtree keeps a running synchronization on both sizes, that the mtree will create its replicated context on the other datadaomain, but that one will be empty. Is there danger that this could potentially remove data from the existing storage unit?
not sure i follow you, are you afraid that it will replication from target to source and delete your data ? When you create replication context you specify source and target DD.
mtree replication is unidirectional from a source DD mtree path to a destination DD mtree path. the destination is in Read Only mode and can not write to it unless replication is broken.
And if you broke replication, what would be the process of mount the destination DD mtree in a device in Networker for restore purpose in the BRS site?
1-Create a storage unit for the mtree
2-Create a device in Networker with the wizard without labeling the device
3.-Umount the volume of the primary DD (volume.001) with Networker Console
4.-Mount the volume in the BRS DD (volume.001) with Networker Console
As Networker have the volume.001 in its database, will Networker know the volume contents?
Is there any adittional step to follow?