Failed to fail-over in VC

[rosseba]

Hi.

My scenario:

1 cluster with 2 ESX 3.0.1, HA and DRS enabled over it, VC ver 2.0.1 build 40644.

Under Task & Events there's every 2 minutes this warning:

Failed to fail-over <myvm> on <esxnode> in cluster <mycluster>

myvm is in a local storage (not a shared storage), but is:

• disabled as HA under mycluster -> VMware HA -> Virtual MAchine Options

• disabled as DRS under mycluster -> VMware DRS -> Virtual MAchine Options

How can I remove this warning?

Best regards.

-f

[rosseba]

More info about the problem:

I set Manual value for Automation Level for the vm and warning stops in VIC Events... Is it correct?!

But now I've another problem: if I stop the vm, then vm poweron automatically after some minutes... but Restart Priority is set to Disabled for this vm...

Can I solve the trouble?!

Regards,

-f

[kjb007]

This behavior is similar to what is seen on the windows services side, so it does not surprise me, but it is new. Just as in Windows, if you disable a service,it will give you a warning, but if it is set to manual, then you don't get a warning. You'll have to get VMware to tell you for sure, but from what I see, if you set it to disabled, then DRS will still try to do what it is configured for, being as it is disabled, it will fail, and warning is generated. It's almost as if disabling does not tell DRS that you don't want anything to do with DRS. Not very technical, but it's just what I think is going on.

The restart priority is with HA, and should have nothing to do with when you stop the vm. Check your vm options, and make sure the stop button is set to power off, and not reset. Also, check the tasks&events on the vm object to see what is exactly occurring.

-KjB

[rosseba]

The restart priority is with HA, and should have nothing to do with when you stop the vm. Check your vm options, and make sure the stop button is set to power off, and not reset.

Stop button is correctly set to power off.

Also, check the tasks&events on the vm object to see what is exactly occurring.

I've the event for Virtual Machine on xxx is powered off and after some minutes Virtual machine on xxx is powered on.

[kjb007]

On the powered on event, does it have an ID associated?

-KjB

[rosseba]

> On the powered on event, does it have an ID associated?

In VIC, I can see only a user Administrator for Virtual machine is stopping and Virtual Machine is powered off, and no user and no task for Virtual machine is powered on next event.

Have I answer to your question? ?:|

-f

[Troy_Clavell]

Can I ask you a question? Didn't you say you are using local storage with HA and DRS enabled? If you have local storage setup in your enviornment, why are you enabling HA and DRS? You need to have shared storage for HA and DRS to work.

Try disabling HA and DRS and you should see the error message go away.

[rosseba]

Hi Troy.

Yep, I know that shared storage is required for use HA and DRS.

BTW, vm with warning reported in events is just a test vm, and it uses local storage.

But,

why are there fail-over warning on the vm meanwhile HA and DRS are disabled for this one?!

How could I modify this environment for support HA and DRS for production vm and local storage vm only, without HA or DRS, for tests and demo vm?!

Sorry but actually I cannot disable HA and DRS on production environment.

Regards.

[Troy_Clavell]

I gotcha... I assumed your entire enviornment was running on local storage. The one test VM that is running on local storage, within the cluster, can you disable the automation level within DRS for that particular vm?

[rosseba]

No Troy,

I've only ONE environment: a cluster with 2 ESX node with some datastore shared from SAN.

This vm is actually on a local datastore in the second ESX node, and because its only a test vm its run locally.

With only one cluster, can I have this mixed situation working or need I to do some changes to the environment?!

[Troy_Clavell]

I guess i'm not understanding you and your are not understanding me. You have a cluster with 2 ESX hosts running on shared storage correct? Within that cluster of 2 ESX hosts, you have one VM that is running on local disk on one of the two ESX Hosts, correct? If I understand correctly, that one HOST that has both shared VM's and a local VM, the host is in a the cluster, correct? So, if the host is in the cluster, by default that makes the VM is part of the DRS pool, because it's linked to the host. So, my suggestion was to edit the settings of DRS under virtual machine options, and change the local VM's automation level to disabled.

If I am wrong in my assumptions and I am still not understanding what you are trying to accomplish, I appoligize.

[rosseba]

> I guess i'm not understanding you and your are not understanding me.

Yep...

> You have a cluster with 2 ESX hosts running on shared storage correct?

Correct.

> Within that cluster of 2 ESX hosts, you have one VM that is running on

> local disk on one of the two ESX Hosts, correct?

Correct.

> If I understand correctly, that one HOST that has both shared VM's and a local VM,

> the host is in a the cluster, correct?

Correct.

> So, if the host is in the cluster, by default that makes the VM is part

> of the DRS pool, because it's linked to the host. So, my suggestion was

> to edit the settings of DRS under virtual machine options, and change the

> local VM's automation level to disabled.

When you tell 'edit the settings of DRS under virtual machine options', do you refer to cluster properties -> edit settings -> VMware DRS Virtual Machine Options?

Because:

• if I set 'Automation Level' = 'Disabled' I've a tons of warnings;

• if I set 'Automation Level' = 'Manual' there aren't warnings.

Why?

Is somewhere another place where set DRS settings?!?!?

[Troy_Clavell]

When you tell 'edit the settings of DRS under virtual machine options', do you refer to cluster properties -> edit settings -> VMware DRS Virtual Machine Options?

Because:

• if I set 'Automation Level' = 'Disabled' I've a tons of warnings;

• if I set 'Automation Level' = 'Manual' there aren't warnings.

Why?

Is somewhere another place where set DRS settings?!?!?

Yes, you are in the right place. Have you tried disabling DRS in the cluster, then enabling it again. I don't understand why you would be getting errors when the DRS rules for that VM are disabled, but none when the are manual.

You may have open an SR with VMware on this one.

[kjb007 · Accepted solution]

If there is no event, then the host is doing it itself. Try re-installing the tools, as the scripts seem to be not working properly. It definitely still looks like you're doing a reset, and not a power-off. Do you have startup enabled in the Startup/Shutdown in the configuration tab of your ESX host? This shouldn't be doing anything, but just a verification step.

-KjB

[rosseba]

> If there is no event, then the host is doing it itself. Try re-installing the tools

Hi Kanuj,

re-install VMwareTools was resolutive for this case.

Furthermore, now if I shutdown the vm, it still down.

Thank you for all tips.

Regards,

-f