VMware NFS Datastores Snapshot Management
Now that VMware NFS datastores are supported on FlashArray through FA File, snapshot workflows with vSphere environments are worth covering in some detail.
Requirements
- FA File for FlashArray enabled
- Purity version 6.4.2+
Recommended reading
Introduction
In this article, snapshot management and recovery workflows will be covered for NFS datastores on FlashArray. There are many options available in vSphere but this guide will cover how to use FlashArray's lightweight and portable snapshots for FA File to recover a virtual machine (VM) in a vSphere environment with one of the VAAI optimized workflows.
Snapshot policies for FA File are used to create and retain snapshots on a schedule that the array administrator determines and configures. These are similar in concept to Protection Group snapshot schedules for block-backed volumes on FlashArray.
In order to have VM granular snapshots, it is recommended to enable automatic directory management on the FlashArray.
If File snapshot policies are already configured on the FlashArray, skip the FlashArray File Snapshot Policies section. If the VM will be recovered from a manually created File snapshot, skip the FlashArray File Snapshot Policies section.
- FlashArray File Snapshot Policies
-
1. To create a snapshot policy on FlashArray for File, log in to the GUI and navigate to (1) Protection, then (2) Policies, then left-click the (3) + sign.
2. Give it a (1) Name and optionally select a Pod for it to be part of then left-click (2) Create. If no pod is needed, the default of none will be sufficient.
3. Left-click the (1) hyperlink name under the Snapshot Policy section that was created in the previous step.
4. Under the Members section, left-click the (1) ellipses and left-click (2) Add Member...
5. (1) Select existing directories from existing file systems and left-click (2) Add. Note that root directories can also be selected here if the file system doesn't have anything else populated yet.
6. Now that there is a directory member as part of this snapshot policy, create a rule to create snapshots on a schedule automatically.
Left-click the (1) + sign under the Rules heading.
7. In the window that pops up, populate (1) frequency of snapshot creation, (2) retention of those snapshots and (3) a Client Name for easier identification of the snapshot origin. Optionally, an additional suffix can be added if it suits the needs of the environment.
8. Reviewing the page that populates here is a good idea to validate the snapshot policy has the correct (1) Members and (2) Rules that were recently configured.
- Recovering From a FlashArray Snapshot in vSphere
-
1. Log into the GUI of the FlashArray and navigate to (1) Protection then (2) Snapshots and determine the desired restore time by reviewing the time that the snapshots were Created. Once the desired restore point is determined, copy the Name for use later in vSphere. The client name (vmwaresnap) and suffix (590) will be useful later in step 8.
2. Navigate to (1) Storage then (2) File Systems and left click on the (3) file system's name of the directory where that snapshot lives.
3. Note the path of the (1) directory where the VM to be recovered as well as the name of the (2) directory export and the (3) path of that directory export on this file system. In this example, the path prefix is /plugin-test-nfs-datastore-01/nfs-load-vm-01/. Copy this path prefix for use in step 8.
4. Navigate to (1) Settings then (2) Network and find the (3) File virtual interface (VIF) IP address. Copy this down for use later.
5. There are a lot of options from this point for how to recover from this snapshot in vSphere. This guide will cover how to recover from a snapshot using VAAI to greatly increase the efficiency of moving files around on the FlashArray once the VAAI plugin is generally available. At the completion of these steps, there will be a read-only NFS datastore mount to the snapshot directory that is being recovered from on the FlashArray.
Log into the vSphere GUI. In the Hosts and Clusters view or (1) Datastore view, right-click on the (2) Datacenter (or Cluster or ESXi host) that should house the temporary NFS mapping that is about to be created to allow for a read-only mount of the snapshot directory from earlier. Highlight (3) Storage then left click (4) New Datastore...
6. Select (1) NFS and left click (2) NEXT.
7. Select (1) NFS 3 and left click (2) NEXT.
8. This screen is where the File VIF address, directory Name and paths from earlier will be helpful. Populate a (1) Name for the datastore.
For the (2) Folder field, use the path prefix from step 3 from above (/plugin-test-nfs-datastore-01/nfs-load-vm-01/ in our example) combined with a path suffix. For our example, the path suffix includes the .snapshot directory, the client name configured in the snapshot policy and a suffix from step 1.
What does this look like when it is all put together?
/plugin-test-nfs-datastore-01/nfs-load-vm-01/.snapshot/vmwaresnap.590
After this path has been constructed, paste it into the (2) Folder field.
For the (3) Server field populate the File VIF address from step 4 and check the box (4) Mount NFS as read-only and left click (5) NEXT.
9. Select (1) the host(s) this NFS datastore should be accessible to and left click (2) NEXT.
10. Review the details and assuming everything looks as expected, left click (1) FINISH.
11. After the NFS datastore has successfully connected to the host objects, left click the (1) datastore name, left click the (2) Files tab, (3) check the box(es) to select the files needed for recovery (in this case we are doing all of them), and left click (4) COPY TO.
12. Select the (1) datastore that's the destination for the VM to be recovered and optionally select a (2) folder on that datastore where the VM's files should live. In this example there is a previously created a folder with the nfs-load-vm-01-restore name so it would be easier to find. It is a good idea to have a folder name that matches the future restored VM's name. Left click (3) OK to proceed.
13. Navigate to the path of the folder where the snapshot VM's files were moved to. Check the box to select the (1) .vmx file and left click (2) REGISTER VM to add the VM to inventory in vSphere.
14. Populate a unique (1) virtual machine name, select a (2) location for the VM, then left click (3) NEXT.
15. Select a (1) compute resource then left click (2) NEXT.
16. Review the operation parameters and left click (1) FINISH if everything looks good.
17. If the VM this snapshot came from is currently powered on and uses a static IP address, adjust the IP address on the source or target VM to avoid duplicate IP addresses or disable the network adapter on one of the VMs. The newly registered VM should power on without issue at this point.