Site Recovery Manager: It’s Not Just for VMs

-

MultiStore-M.jpg

A key goal at NetApp is advancing the value of storage in the virtual data center from the traditional model of highly available storage to a more advanced virtual object oriented model. This approach is unified in nature as it provides the traditional functions of high availability, high performance, and the traditional storage provisioning of LUNs and file shares while also expanding into the realm of management of virtualized objects by the virtualization administrative team.

By virtualized objects I am specifically speaking of the objects that a VMware admin care about such as VMs, Datastores, workflow automation tools such SRM or View Manager, etc. For this blog I’d like to share with you an example of the direction NetApp is taking storage virtualization by discussing how to integrate a NetApp NAS file server into a VMware SRM recovery plan.


NetApp Virtual Filers

In 2002 with the release of Data ONTAP 6.2, NetApp introduced MultiStore and the Virtual Filer.

With MultiStore customers are able to logically partition a physical NetApp array into multiple virtual storage arrays referred to as Virtual Filers. A vFiler is an isolated software container that behaves exactly like a physical NetApp array. vFilers share the physical resources of the array and by abstracting the client access from physical array into virtual arrays.

Each vFiler is its own unique entity complete with its a hostname, IP address, iSCSI Qualified name, and Active Directory SID or Kerberos realm ID, etc. For all implicit purposes vFilers are just like VMware VMs, and as such they provide customers with similar advantages over physical storage arrays.

vFilers provide the ability to non-disruptively migrate a production dataset between NetApp controllers in a HA configuration. vFiler migrate is analogous to VMware’s VMotion technology. vFilers can also be pre-configured for recovery in DR scenarios. This functionality is referred to as vFiler DR, it is analogous to VMware’s Site Recovery Manger, and it is the topic I am going to discuss in this post.

If you’d like more details around MultiStore I’d recommend you see Scott Lowe’s recent post on MultiStore.

Virtual Filers sound a lot like Virtual Machines. Marketing Ploy?

I’m sure some will make this claim; it seems to be the reaction some take to the unique virtualization technologies that NetApp ship. In fact I’m having premonitions of such responses as I type!

The reality is NetApp has a long history of delivering unique storage virtualization of which vFilers we added to the portfolio in 2002.

Integration is Simplification

I think we’d be hard pressed to find someone who wasn’t enamored with the elegance SRM delivers to the numerous tasks associated with implementing a DR plan to recover virtual machines. This simplicity has made what was once only available to an exclusive list of applications a reality for every virtualized server.

Along with the recovery of your virtual machines, properly recovering a whole environment might require the recovery of NAS storage as well like CIFS shares for home directories or engineering datasets. For example if you’ve recovered your virtual MS Exchange server with SRM and your users keep their Outlook profiles or .pst files in their home directories they’ll need access to their home directory as well.

By integrating the execution of the vFiler DR command into a SRM recovery plan you can automate the recovery of your vFiler, extending the simplicity and value of SRM should a worst-case scenario ever occur.

vFiler DR Details

Any time after a vFiler is created it can be configured to failover to a remote NetApp FAS or V-Series array. This configuration is completed on the destination NetApp array buy running the vFiler DR configure command. For the sake of this post I wont go thru NetApp setup steps as they are well documented.

vFiler SRM Pre-M.jpg

Unlike VMs, VFs do not require the equivalent of vCenter Server in order to implement vFiler Migrate and vFiler DR as Data ONTAP executes these processes. When a vFiler DR activate command is issued the running vFiler is shutdown (if there is a running vFiler), the SnapMirror replication is broken, the datasets are brought online, and the vFiler starts. The process of bringing a vFiler online can be preconfigured to start with a new IP address, which gets registered with DDNS servers. To anyone familiar with the capabilities of SRM this feature should sound very familiar.

vFiler SRM Post-M.jpg

How to integrate vFilers with SRM

I have to tip my hat here to VMware as they have made putting this solution together pretty straight forward by allowing recovery plans to execute additional commands. As you can see here in this image that we are simply adding a script that will remotely execute the vFiler DR command.

srm_command_3-Zoom.jpg

A critical note: As a safety measure the vFiler DR activate command shuts down the production vFiler prior to starting the DR version of the vFiler, this design prevents concurrently running two instances of the same VF. Because of this behavior I would advocate the insertion of a confirmation step into vFiler DR activate scripts prior to the execution of the activation step. Doing so would allow the running of a SRM recovery test without forcing the execution of vFiler DR, which would disrupt the production vFiler environments

Want More?

I am mulling around posting a follow up to this blog where I cover the exact steps of configuring the vFiler DR setup, securing remote execution of the script, and sharing an example script. Let me know if this level of detail is of interest and I will start on the follow up, else I’ll move on to the next post.

I’d also like to thank our resident SRM expert Larry Touchette, author of TR-3671: VMware Site Recovery Manager in a NetApp Environment, for his feedback and assistance on this post.

Vaughn Stewart
Vaughn Stewarthttp://twitter.com/vStewed
Vaughn is a VP of Systems Engineering at VAST Data. He helps organizations capitalize on what’s possible from VAST’s Universal Storage in a multitude of environments including A.I. & deep learning, data analytics, animation & VFX, media & broadcast, health & life sciences, data protection, etc. He spent 23 years in various leadership roles at Pure Storage and NetApp, and has been awarded a U.S. patent. Vaughn strives to simplify the technically complex and advocates thinking outside the box. You can find his perspective online at vaughnstewart.com and in print; he’s coauthored multiple books including “Virtualization Changes Everything: Storage Strategies for VMware vSphere & Cloud Computing“.

Share this article

Recent posts

10 Comments

  1. Thanks for the link back to my article!
    I can’t speak for everyone, but personally I’d love to see a more in-depth follow-up to this post with more details on the script and the vFiler DR setup.

  2. I’d love to see the script and setup detailed. This is a game changer as it unifies the data protection with protection of the Virtual Machine.

  3. Any guesses as to when vFilers can be fully managed from a GUI? (whether FilerView or Storage Manager)
    For us, that’s been one of the larger drawbacks so far with MultiStore.
    Thanks for the good post.

  4. I would really appreciate a follow-up article with more details about the necessary steps required for vFiler DR in combination with SRM.
    thanks a lot!

  5. If you can’t incorproate vFiler DR into tests of the recovery plan, how can you be sure it will work as expected in a real recovery?
    I remember, back in the days when I was actually responsible for the DR plans for multi-million dollar systems, I learnt the hard way that a plan is worthless if it’s never been fully tested.
    So, I wouldn’t be comfortable using a feature for DR that I couldnt incorproate into my DR tests…

  6. Certainly you can incorporate vFiler DR into your tests of a recovery plan. Once the DR vFiler site is configured, it is a matter of issuing a single command at the DR site (which is what the script described in this article does), which activates the DR vfiler. Perform your test. When your test is complete, issue the “vfiler dr resync” command to synchronizes the DR and production site, and then bring up the production vfiler again.

  7. “and then bring up the production vfiler again.”
    so if I understand correctly, you’re suggesting either:
    – Test vFiler DR by actually failing over a vFiler to the DR site, meaning the vFiler (and whatever it’s serving out) is not available to things still running in the production site for the duration of the test.
    or
    – Just hope that the vFiler DR operation will work when you actually need to failover.
    How about a third option – just use some windows VMs to serve out those “shares for home directories or engineering datasets.” and get all the benefits of Site Recovery Manager (i.e. Zero disruption to operations in the primary site during testing AND assured recoverability)?

Leave a Reply

Recent comments

Jens Melhede (Violin Memory) on I’m Joining the Flash Revolution. How about You?
Andris Masengi Indonesia on One Week with an iPad
Just A Storage Guy on Myth Busting: Storage Guarantees
Stavstud on One Week with an iPad
Packetboy on One Week with an iPad
Andrew Mitchell on One Week with an iPad
Keith Norbie on One Week with an iPad
Adrian Simays on EMC Benchmarking Shenanigans
Brian on Welcome…