This discussion is now open for questions to the experts. we look forward to an interactive and vibrant discussion on the topic.

There are a couple of common issues we see with using replicated data on a Data Domain system for a DR Test.

The first is that the replicated copy is read-only, and is still actively being replicated to (unless you stop replication for the test).  This can sometimes cause problems for your backup application if it is expecting the data to be in a read-write state.  And what can really trip up your DR test is if the data is moved/deleted on the source system, which then gets replicated to the destination/DR box.  Getting "Data not available" or "data not found" messages during a recovery exercise is not a good feeling, which is why the best practice is to make a copy of the replica data, either via fastcopy or a snapshot, to use for the DR Test.

This leads to the second issue we commonly see:  A frequent call to Data Domain Support involves utilization on the replication target system being higher than on the source system.  On investigation, the Support Engineer will often find there are more files on the target system, and can often identify the date range of those files by looking at the "File Distribution" section of an Autosupport.  Their next question is usually, "Did you do a DR Test around ZZZZZ time?"  Adding a step to your DR Plan to cleanup the copy of the data used for testing will prevent this scenario (but should obviously be skipped when executing a true Disaster Recovery).

Hello,

I was curious if you see a lot of customers implement snapshots on target DD unit ?  Let's say i have active replication and DBA accidentally deleted an Oracle archive file, a few seconds later it is purged on target DD as well. If i had snapshots on DD (say daily), would i be able to recover that file ?  Can i even export a read-only snapshot from DD via NFS/CIFS ?

Thanks

Hello Scote,

what i mean in my example was a DBA using data pump (expdp) to create exports on Data Domain for backup purposes. Currently Data Domain is locally mounted on Oracle server (NFS) so DBAs have direct access to their NFS export and run their own "clean up" scripts.

I'm investigating using 2 DD appliances to replace tape on 2 locations (using Replicator).

Where can I find a best practices document on implementing and testing DR?

Thanks.

Hi Mark

Currently and after collection replication; all the virtual tapes in DR Data Domain are replicated under vault but in Main Data domain those tapes are under tapes not vault, and all the tape state in DR DD in RO/RD and of-course the tape state in main DD are in RW; so whats is RD state and how can we change the state of tapes in DR to RW in disaster scenario? and keep in mind that main data domain is configured as storage library on main Networker environment and the DR DD configured on DR Networker environment and once we failover Networker we will need to change the state of DR data domain tapes to RW

Hello Ahmed

This has been already answered. Basically the data on the destination of a replication pair, is in read-only mode.

You have to break replication so the data in destination is r/w.

Hi All

We are trying to implement DD-Boost -Oracle and we have a question regarding secondary copy (DR copy)

Scenario-

Lets say We have server name  oracle1 backup to -DD1 using  RMAN replication file system we copy that data to DD2 so oracel1 knows about copy 2

can we do a refresh of that data from DD2 to different oracle server located at DD2 site that is lets say Oracle2 is that possible if so how tried to find documentation did not see any.

What is the best replication options is RMAN file replication or Mtree replication is best we do refresh DB at DR site a lot than whole DR scenario.

Thank you.

As RMAN is fully aware of copy, key question is which copy would RMAN use to restore?  This does not depend on DD, but rather RMAN.  As such replication copy would be RMAN aware and part of catalog, I assume there should be a way to say use replication copy when using RMAN alone.

Both replication options are valid (see https://www.emc.com/collateral/hardware/white-papers/h10683-dd-boost-oracle-rman-tech-review-wp.pdf), but I would argue best one is the one which gives fuller integration with application options.

So if i understood this correctly Mtree replication also can help in refresh from DD2 but did not find any documentation that DD Mtree replication copy is aware by RMAN so got confused if that replication is option or not

I think by  Mtree replication RMAN will nto have information about secondary copy(Correct me if i am wrong) if i have to do refresh from that copy how can i do is my biggest issue and question from Oracle DBA

Not sure how you can control that.  I use backup application to control that.  Without backup application, it should env variable coming with ddboost plugin which should control that (if possible), but as I don't use that approach I do not know all possibilities.

You need separate thread for your question as this ATE session ended I believe.. 

Hi,

I have a couple of questions:

1. Can user roles be combined in order to create a customer role in a Data Domain?

2. Can Data Domain integrate with 3rd party management platforms such as Microsoft Sysyem Centre, via management packs?

3. Is there a roadmap for 16Gb FC and infiband support on the DD4200?

These are question for an RFP response, any help will be much appreciated.

Regards

Ron