spinning up a vdb from an incremental change and replicate

  • 0
  • 1
  • Idea
  • Updated 2 weeks ago

Hi, Is there a possibility that I can spin up a vdb from just an incremental change/backup or the delta change between snapshots to then replicate it using 'selective data distribution'.  If this can be done the question of subsetting is no more.
OR
I can put it this way, is there a possibility to segregate a vdb after provisioning from dsource to then replicate it to functional test environments
Regards.
Photo of Sri Ramasamy

Sri Ramasamy

  • 160 Points 100 badge 2x thumb

Posted 2 weeks ago

  • 0
  • 1
Photo of Anders Karlsson

Anders Karlsson

  • 170 Points 100 badge 2x thumb
Hi Sri,

Thanks for your question. I'm not fully following the use case here. You might be able to help me by explaining why the current method is not working. 

The dSource is the endpoint which is taking the backups and is responsible for provisioning the change to the vDBs. It is all working on deltas, including SDD which has a specific set of 'instructions' to ensure sensitive data is not replicated. 

Kind regards, 

Anders
Photo of Sri Ramasamy

Sri Ramasamy

  • 160 Points 100 badge 2x thumb
hi Andres,

so the objective is to have small DBs for functional testing not the production volume or a vDB provisioned from a snapshot which when replicated will bring in dSource/prod volume.

the current method works, no issues but it will always provision a full DB.

we've have 'selective data distribution' where I can replicate just the masked vDBs, which again is a full size DB.

So how I can bring in say thousand rows from a 10TB DB? subsetting??, which is a complex process, but we can still think of hooking Application Data Model (Oracle's Subsetting) to Delphix.

Is there anything else, yes, if I can spin up a DB from just an incremental backup holding just that delta size.. that would be amazing (please think beyond Oracle)

the other option would be create a 'masked' (labelled as masked using Demo job but nothing is actually masked, selective data distribution will only replicate(without dSource) a vDB if it's labelled as masked), vDB then run some rule set that will update/mask say 1000 rows. then replicate the just the blocks private to this vDB?

Delphix solution if designed for developers where they want to have many functional environments, early testing/performance environments should provide this flexibility..  I think... may be in 5.4?

Regards.
Photo of Sean Nothdurft

Sean Nothdurft, Technical Support

  • 1,000 Points 1k badge 2x thumb
Hi Sri-

The Delphix Virtualization Engine currently only provides 1:1 copy of the source database in the VDB provision process, with masking or other transformation operations available within the product. We do have functionality requests to support provision of a subset of Source data, but we currently have no ETA for this kind of function change or enhancement. We can certainly add your organization to this request to indicate interest.  

Otherwise, to accomplish your request you may consider hook scripts to drop a significant amount of data from the VDB provisioned. Our documentation may help to get you started, but for detailed assistance our Professional Services team may be engaged to help further.
Photo of Sri Ramasamy

Sri Ramasamy

  • 160 Points 100 badge 2x thumb
thanks!
we've considered the option of including hook scripts but with our massive DBs holding some >100k partitions/subpartions per schema and the data of a schema distributed across tablespaces also with virtual columns makes it complex and time consuming. The slicing down of a DB takes days. 

we'd happy to see such features in future.. 

regards
Photo of Sean Nothdurft

Sean Nothdurft, Technical Support

  • 1,000 Points 1k badge 2x thumb
Understood - thanks so much for your feedback, and please let us know if we can assist in any other way!  As mentioned, we will add your organization to the outstanding request for future enhancements.

Take care!