ESX VM failback?

I had a power outage on one of our ESX 4.0 hosts last night. We run a 2 ESX host cluster via VCenter so all those Virtual Machines shifted to the other ESX host. What was interesting is when power came back on 4 minutes later, none of the Virtual Machines failed back to the other server as we had them for load balancing.

I checked the configuration tab and under networking I selected the VSwitch with the VMs on it and the little weird box to the left when clicked did say Failback enabled on both ESX hosts.

How do I get failback working and is this actually a good idea or is it better to go in and manually move the VMs back?

Thanks in advance guys/gals
Who is Participating?
bgoeringConnect With a Mentor Commented:
I think that you may be talking about the failback on NIC Teaming tab? If so that setting has effect when you have standby NICs configured on the vSwitch so that if your main NIC or NICs go down, the standby NIC takes over. The NIC will go down in the event the physical swith it is plugged into goes down, someone trips over the cable, etc. Anyway when the primary becomes available the network traffic will fialback to the primary NIC.

In HA, it sounds like it is woking as designed. When one host went down it restarted your vms that were running on that host on the remaining host.

There is no "failback" setting for moving the VMs back to the original host when it becomes available. However, if you have DRS enabled on the cluster it will move hosts back if the load is too high on the 2nd host, and continue to vMotion hosts between the cluster Nodes whenever the host utilization reaches a threshold that would cause DRS to decide to balance the load.

There are DRS rules you can create that will dictate the DRS keep the VMs running on seperate hosts -- doing that MAY be able to give you the results you want. But not sure if trying that would prevent HA from restarting them. Hmmm.

If you want specific VMs running on specific hosts, you will need to manually move them.

Hope this helps - Good Luck
StappmeyerConnect With a Mentor Commented:
Keep in mind that even with DRS enabled, if the load is not high enough on the host with all the VMs, it will not vMotion guests over to the server that had the power outage and then came back into service.  Ensure that you have DRS enabled and ensure that you have set how aggressive you would like it to apply recommendations.   These setting along with the actual VM Guest requirements will cause the loads to be shifted.

Paul SolovyovskyConnect With a Mentor Senior IT AdvisorCommented:
If are you doing DRS it will not "failback" unless there is lack of resources.  If your single system does not run into resource issues all the VMs will stay on there unless you're going a more aggressive DRS configuration.
Keep up with what's happening at Experts Exchange!

Sign up to receive Decoded, a new monthly digest with product updates, feature release info, continuing education opportunities, and more.

snyderkvAuthor Commented:
No we can have all VMs on one server no problem. Only all the VMs shift over to the second ESX host every time their is a power failure on one or the other. If this happens every day, then you can see how much of a pain it would be to manually migrate half those VMs over.

Yes I believe DRS is enabled. and I see the Failback option but I think that was something else like what bgoering mentioned
Sorry I don't have better news for you. How do you get power failure on only one server if both are in same datacenter? Only UPS for one? (Just curious)
snyderkvAuthor Commented:
This case may have been one of the remote site admins tripping over the cable or mistaking it for something else.

At least I know the answer to my question. So thanks again everyone.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.