EMC RecoverPoint- Standalone VM Recovery

RecoverPoint is one of my favourite products. It’s unique, reliable and has helped me out of trouble on more than one occasion.

Whilst most VM admins associate it with SRM and thus Site/LUN/Datastore recovery, it can also fill in an often overlooked gap in data protection. Pre and Post backup coverage. The ability to recover a vm (or data) in the window before or after your backups run.

Depending on your configuration, you’ll likely have journals available to recover from every few seconds for the duration of your Recovery Window. If a failure were to occur after your last backup, but before your next, this can provide another “safety net” to have available. Obviously this is determined by a a number of conditions: Link bandwidth, Data Rate of Change, Latency etc.

Start by logging into Unisphere for RecoverPoint, then click on the “Protection” tab.

image1new

 

 

 

 

 

 

 

 

 

 

image1a

Note that the Copy & Group Image access actions controls are greyed out. This is due to the Consistency Groups being managed by SRM by default. To allow the Image Action controls to be used, go to the “Policy” tab for the Consistency Group being used, and then into the “Stretch Cluster / VMware SRM Support”.

image3

 

 

 

 

 

 

 

 

 

Click on the “Group Policy” Tab, and change the “External Management, Managed By:” option to “RecoverPoint” & click Apply, then close the Success dialog. You could do this to more Consistency Groups using RPCLI/Powershell as I’ve shown in this link.

image4

 

 

 

 

 

 

 

 

 

Return to the “Status” tab, expand the CG and highlight the “CRR” and the action buttons will be available.  Click on the “Test Copy” button.  The Wizard will start. (This was formerly known as Enable Image Access).

image5

 

 

 

 

 

 

 

 

 

 

 

We want to pick a specific Image/Bookmark/Journal from the list so click the arrow and choose “An image from the list”. The list of available images will be displayed. The amount of images will depend on the applicable Protection Window available to the LUN. Choose an image

image6

  • The most suitable access mode for testing is the default “Logged Access”. This requires less steps and best performance, at the expense of being slower to configure during testing. It can then be copied/cloned off to another datastore.
  • Click on “Next Test the Image” and wait while the system processes the test

image13ptii

 

 

 

 

 

 

 

 

 

Now you need to logon to your Virtual Centre host, select a host, and under configuration, storage adaptors, select an adaptor and select Rescan All.

The Replica LUN will be shown as a “snap” Datastore that can be browsed.  Browse the Datastore, find the folder for the VM’s VMX you want to test and add it to the VM Inventory, labelling it clearly. If you’re testing, bring it up in an isolated network to avoid conflicts.

image7new

image8

After you start the VM, when prompted, choose “I copied it” to prevent creating duplicate uuid’s.

image9

The VM will start and can be tested or copied/cloned for Recovery.

zzz

 

 

 

 

 

 

 

 

 

 

Keep an eye on the Image Access log capacity, ensuring it doesn’t fill up.

image11

 

  When you’re finished, shutdown any VM’s on the Replica LUN and remove them from the VM Inventory. Don’t ‘delete file from disk’, but unmount the Datastore.

image12new

From the RecoverPoint GUI, choose the “Cancel” option and the Image Access will be reverted and the hosts will lose access to the Image.

image13

 

 

 

 

 

 

 

 

image14 image15

Change the “group Policy” for the Consistency Group back to “External application” to allow SRM to control the operations. (Normal).

The process is now complete. Now you have another easy, reliable option in your recovery toolkit.

 

 

 

 

Related Post

9 thoughts on “EMC RecoverPoint- Standalone VM Recovery

  1. couldn’t agree more about your feedback related to recoverpoint. couldn’t help but notice your VMFS sizing, would be curious to hear your approach, seems to be variable sizes.

    1. Hi Duane, yes, that one is a bit scattered. A mix of licensing, connection, political and environment “challenges”. Sometimes “you have to do what you can with what you’ve got” 🙂

  2. couldn’t agree more about your feedback related to recoverpoint. couldn’t help but notice your VMFS sizing, would be curious to hear your approach, seems to be variable sizes.

    1. Hi Duane, yes, that one is a bit scattered. A mix of licensing, connection, political and environment “challenges”. Sometimes “you have to do what you can with what you’ve got” 🙂

  3. Hi Brett,
    I do not know if it is appropriate but since the post is related to RP and vSphere, I wanted to write few lines on my recent DR experience.
    I am the storage and vSphere admin at an organisation, with two main sites each acting as a production and DR for the other site. One VNX 5500 array in each location and a vSphere cluster composed of a dozen of hosts. And RP/SE to replicate between sites.
    Two weeks ago we had a site failure, by the look of it vSphere cluster lost connection to the storage and in the morning I had to fail over 9 datastores/LUNs to the recovery site. At the DR site I began mounting the datastores, but only three hosts were able to see them the rest were refusing.
    I had to SvMotion VMs from the copy LUNs onto datastores that were mounted to the rest of the hosts, which increased the RTO significantly. Anyway. VMware support identified that the volumes were snapshots (esxcli storage vmfs snapshot lists), and then I had to mount those “snapshots” with the esxcli storage vmfs snapshot mount -n -l “VMFS01” command.
    My question is have you experienced similar situations when copy RecoverPoint LUNs are detected as snapshots by vSphere at the recovery site?
    EMC support say it is a vSphere issue, and VMware Support say it is how the storage presents it to vSphere. From 9 LUNs I had configured in a consistency group that I failed over I was able to present/mount only 7. Two LUNs are not there. There are there (I found their NAAs) but when trying to mount them they are empty volumes and is not asking me to preserve the signature but to format them.

    Brett, apologies for the long post.

    Kind Regards,

    1. Hi Arthur, sorry i missed your comment. Caught in an over-zealous spam filter. Never nice when your DR safety net doesn’t work as you’ve hoped. I haven’t seen what you have experienced. Does the same thing happen when you test a CG copy ? Hard to diagnose from here, and woudl need to know all versions etc, but I suggest you signup and post up in https://community.emc.com/community/products/recoverpoint/content

      It’s easier to engage there and there’s also some really excellent RP engineering guys who will assist. Hope you can 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *

%d bloggers like this: