VMs on same Datastores using different transport method

[rscale]

We have a scenario where we have VMs in the same cluster, on the same datastore, using the same proxy are using 2 different transport methods (SAN and NBD). We're trying to get have them to all use SAN and have verified that the LUNS are presented to the backup proxy.

I've noticed in the logs of the ones that use NBD the following:
<transports seq="nbd" asyncIo="True" nbdCompressionType="none" />

In the ones that use SAN I see:
<transports seq="san;nbd" asyncIo="True" nbdCompressionType="none" />

Why would san be an option on one VM but not another if they are both on the same datastore? Also both are the same OS (Windows 2012) and use the same vmware tools version.

Thanks

Mike

[veremin]

Doesn't this VM reside on vSAN by any chance or have at least one disk located on a VVOL (see this section)? Thanks.

[rscale]

We aren't using vSAN. The datastore is presented from a traditional FiberChannel SAN (3Par).

[Gostev]

transports seq="san;nbd"

This literally means: "Transport modes sequence: try to use SAN, and if it fails then failover to NBD". If you disable failover to NBD option in the backup proxy setting, you will see this string change to have SAN only.

[rscale]

These 2 VMs are being backed up by the same Job, on the same proxy, and reside on the same datastore. Shouldn't they show the same options in Transport mode sequence?

Thanks!

[foggy]

Looks like they should. Haven't you contacted support regarding this?

[rscale]

This issue went away when we migrated the VM that was having issue to a different ESX host in the cluster. Once we moved it, it was able to be backed up again by direct SAN. No other VMs on that host had any issues. For now, It looks like it was a one-off issue. I'll update the thread if we see any additional cases of this. Thanks!

Mike