Multi-Site Setup

Hi Wanaka,

- If I need to deploy VIPR, should I deploy three VIPR instances?

In the described case you need to manage 3 RP pairs which connect our sites as mesh topology. So all resources (RPs, VPLEXs, VNXs, Fabric Managers) have to be managed by a single ViPR instance.

Please consider latency RTT<150ms limitation for such config.

- How will VIPR manage the VNX that is virtualized behind the VPLEX?

You need to install SMI-S providers to manage VNXs. All backend (in perspective of VPLEX) operations will be done directly on VNXs via SMI-S. Next fronted operations are done directly on VPLEX.

- Can I use VIPR to provision a volume and have it replicate to a different country if I have three VIPR instances?

No, single ViPR instance is required.

You can federate several ViPR instances for a single sign-on, and tenants related info replication. But single orchestration action can be managed by a single ViPR instance only.

Just to clarify "why?". Simple volume provisioning in you case will consist on these basic actions:

- create data and journal volumes on Site-A (with SMI-S)

- create / check VPLEX backend zones in SAN on Site-A (via Fabric Manager)

- create Storage Group on VNX to export volumes to VPLEX on Site-A (with SMI-S)

- claim these volumes under VPLEX Local on Site-A (with VPLEX)

- create / check VPLEX frontend zones in SAN on Site-A (via Fabric Manager)

- create masking view on VPLEX to export volumes to RP on Site-A (with VPLEX)

- create data and journal volumes on Site-B

- create / check VPLEX backend zones in SAN on Site-B

- create Storage Group on VNX to export volumes to VPLEX on Site-B

- claim these volumes under VPLEX Local on Site-B

- create / check VPLEX frontend zones in SAN on Site-B

- create masking view on VPLEX to export volumes to RP on Site-B

- enable replication between data volume on Site-A and replica on Site-B

- export data volume to a host

All these actions contain a lot of pre-checks and very dependent on each other. So only one ViPR Controller instance can manage that.

Hi Wanaka,

You just need to deploy a single ViPR Controller instance that will control all VPLEX and RecoverPoint clusters. You cannot use ViPR to create RP replicated volumes between ViPR instances, all sites need to be under control from the same ViPR instance.

Let me know if you have more questions.

Regards

If I can take this discussion one step further I would like to ask, as the purpose for the VPLEX and RP is most likely to insure availability in the case of a disaster, how would one insure that in the event that site A is lost ViPR would be available in either Site B or Site C?

If a new thread is required I will be happy to post a new thread

thank you.

Hi,

With the single instance and let said I can fullfill the 150ms rtt across the WAN to the other two countries, how many projects should I setup in VIPR?   Should I setup project A for country A, project B for country B and project C for country C?   All three countries are managed by the same storage administrators.   The reason I am asking this is because I will have production running on Country A and having it replicated via RPA to country B or vice versa.   I will need to have SMI-S provider setup in each country.