External Datastores | Nutanix Community
Skip to main content
Solved

External Datastores

  • November 2, 2021
  • 2 replies
  • 365 views

Cantoris

Hi!

We have two Nutanix clusters running ESXi v6.7.0 with a Horizon 7 VDI environment on each.  We needed to be able to sync Horizon AppVolumes between the clusters via some common storage.

The solution we were set up with was a FreeNAS VM on one cluster publishing an NFS share which was then attached to the hosts in both clusters.  We weren’t bothered about high availability as it’s not an issue if the replication of the appvolumes goes down sometimes as it’s only needed if appvolumes are being changed or their assignments to users is changing.

After an NCC upgrade, Prism Element now flags up this external datastore as unsupported.  I’m struggling to find out from Support what the actual magnitude of risk is here to the stability of the clusters (especially when it comes to upgrading firmwares, AOS etc).

Does anyone here have any thoughts on this situation please?

Thanks for reading!

Best answer by Cantoris

Thanks for the reply!

I did mention this to Support but they said it would also fail the NCC check as it would “still use the vmkernel and CVM for the IO”.

There’s actually a solution article for the exact setup and situation we’re in (bottom section): VMware App Volumes in a Multisite Deployment

Apparently what that article suggests would also be flagged up by NCC under this check!

Nutanix have now advised me that they believe the risk in our case to be minimal and to leave as is.

View original
Did this topic help you find an answer to your question?
This topic has been closed for comments

2 replies

JeroenTielen
Forum|alt.badge.img+8
  • Vanguard
  • 1363 replies
  • November 2, 2021

HI @Cantoris,

Cant you create a volume group and attach it to the needed machines?


Cantoris
  • Author
  • Voyager
  • 2 replies
  • Answer
  • November 5, 2021

Thanks for the reply!

I did mention this to Support but they said it would also fail the NCC check as it would “still use the vmkernel and CVM for the IO”.

There’s actually a solution article for the exact setup and situation we’re in (bottom section): VMware App Volumes in a Multisite Deployment

Apparently what that article suggests would also be flagged up by NCC under this check!

Nutanix have now advised me that they believe the risk in our case to be minimal and to leave as is.