VDP 6.0.2 and DataDomain

[JPZirkle01]

I am currently testing VDP 6 performance and replication for possible full deployment. In my testing, I've run into an issue that I cannot find information about. My general configuration is below:

Site 1

VDP 6 Appliance with datastore backed storage.

Runs full backups of guests.

Site 2

VDP 6 Appliance with small datastore backed storage AND DDBoost access to DD2500.

Intent is to be replication target of Site 1 VDP.

Site 2 VDP with DD functionality works fine for backups initiated on Site 2 VDP, because you select the DD as the target for storing backups.

My issue arises when I attempt to configure a replication job on Site 1 VDP making it's target Site 2 VDP. I am able to establish the connection properly and replication works properly. The caveat is that it will ONLY write to the direct, DS backed storage and NOT to the DD.

Has anyone attempted this and found a solution? I found no documentation on this, nor any options to configure the replication to use DD instead of the "standard" storage.

Appreciate any feedback.

[ASRIS]

I believe (could be mistaken) that backups stored on VDP1 (on DataDomain) will be replicated to VDP2's DataDomain.

So if your backup is on VDP1's local disks then it'll save it to VDP2's local disks.  You'll need each VDP will need to connect to their own DataDomains.

[JPZirkle01]

Thanks for the reply.

I had assumed this would be the case, but found no documentation to back it. I do have a DD at each site, however Site 1's DD is at 90% capacity with our conventional backup data. Was trying to get around the DD replication until I can drain the conventional backup data. I will assume that VDP "local" storage will only replicate to VDP "local" storage and VDP via DD will only replicate to VDP via DD.

[jhunter1]

Correct - source and target must be similar. VDP+DD <---> VDP+DD or VDP <---> VDP. VDP uses DD copy to replicate backup data between DD systems. Metadata is stored and replicated between VDP appliances.

[JPZirkle01]

Thanks for the reply, jhunter.

One final question on this scenario. Currently, I'll have to do VDP1 > VDP2 replication. Once DD1 has free space, I intend to reconfigure all jobs on VDP1 to begin using DD1 and reconfigure repl jobs to do the same to allow for VDP1+DD1 > VDP2+DD2. That's straightforward enough.

Is it possible to REMOVE VDP storage without destroying the appliance and having to rebuild? Once I've shifted to VDP1+DD, I no longer need the 10TB of "local" storage attached. If this is not possible, I'll need to adjust my plan to build a NEW VDP appliance in Site 1 to utilize DD1 (with a small VDP configured storage pool). I've found no documentation on permanent removal of "local" VDP storage without destroying the appliance. Thanks again for your feedback.

[RyanJMN]

There is no way to reduce capacity of an appliance so you will have to build new.

[ASRIS]

Once you remove all your backups off the VDP VMDKs just Storage vMotion the appliance to another datastore using thin-provisioning.   This will reduce the physical foot print of the storage usage.

[JPZirkle01]

Excellent work-around, ASRIS. I'm so adverse to thins, I didn't even consider that option. I'll figure out if I want those to linger or not. I'd hate to deal with the gaps in backup retention if I have to build out a new appliance, so I might do this.

Thanks all, considering this answered.