VMware Cloud Community
neil_murphy
Enthusiast
Enthusiast
Jump to solution

FT using crossover cable

Hi all.

I have a situation where a customer wants to implement FT for a single VM running on a 2-node vSphere 6 cluster. The network fabric is 1Gb and FT now requires 10Gb because of the new "fastcheckpoint" synchronisation method for SMP-FT. One possible scenario would be to install a 10Gb NIC in each host and use a crossover cable.

Do you think this would be supported? Would it even work?

Many thanks,

Neil.

1 Solution

Accepted Solutions
neil_murphy
Enthusiast
Enthusiast
Jump to solution

Thanks Matt,

Good to see that the VMware lab guys have done something similar in the past -  that gives me an extra bit of confidence. I take your point about there being some loss of flexibility when it comes to rolling upgrades and maintaining redundancy levels during a host failure and I'll pass these points back to the end user.

Neil.

View solution in original post

Reply
0 Kudos
2 Replies
MattMeyer
VMware Employee
VMware Employee
Jump to solution

This is similar to how the setup was tested for the performance whitepaper on the legacy version of FT that can be found here:

https://www.vmware.com/files/pdf/perf-vsphere-fault_tolerance.pdf

I cannot think of a reason why it wouldn't be supported without a switch in the communication channel.  The configuration for FT on the host is the same.  You may have to come up with some creative ways to upgrade the hosts without a 3rd host.  You will also give up the automatic remediation of FT due to a host failure.  vSphere HA will normally restart a failed VM on another host and reconfigure FT sync for you.  Without another place to run the VM, you'll be left in a reduced redundancy state.

TL;DR: Will work fine and be supported, but you'll lose some things.

neil_murphy
Enthusiast
Enthusiast
Jump to solution

Thanks Matt,

Good to see that the VMware lab guys have done something similar in the past -  that gives me an extra bit of confidence. I take your point about there being some loss of flexibility when it comes to rolling upgrades and maintaining redundancy levels during a host failure and I'll pass these points back to the end user.

Neil.

Reply
0 Kudos