I recently got asked a couple of times how to replicate or mirror writable volumes across multiple pods. Because the VMware documentation is lacking this information I decided to write a small blog post about it.
First of all, writable volumes are not designed to be used on multiple datastores. A writable volume is an active part of the users desktop and it is not wise to have multiple versions across multiple datastores. So creating a storage group and replicate the writable’s from one pod to the other is not a good idea.
Second to that the replication is always slower than a user logging off and in on a virtual desktop.
For example. If a user is logged in on pod 1 and performs a quick logoff and log in, and lands on pod 2. This process can be done within 1 minute. Within that 1 minute the existing copy of the writable on pod 2 is not overwritten by the replication mechanism. The user is able to log in on pod 2, but will get a writable attached with old data. Meanwhile the replication fails because the writable on pod 2 is in use and cannot be overwritten.
The third thing is. If you have copy’s of writable volumes on multiple datastores, which copy will be replicated? If for whatever reason an old one is replicated the user might lose data.
The solution to this is to create a shared datastore which can be accessed by all pods in your environment.
In the App Volumes Manager, you then need to configure the shared datastore as Default Storage Location for your writable volumes. You can do this by selecting the Configuration tab and then select Storage.
This way you will have one writable volume for every user on one location, and the writable can be attached to virtual desktops on both pods.