Custom alarms

[dragoscts]

Hi,

I want to configure some alarms which will send notifications on email but I can't figure it out what type of conditions should I use.

The scenario includes 2 VMs that I want to keep on different hosts from the cluster.

I tried was the following rule:

Targets: All virtual machines in vCenter server servername

Alarm Rules: IF VM connected

Argument vm.name starts with vmname-

Argument host.name equal to first-cluster-name (where both vm runs)

THEN trigger the alarm as  warning and send emails

OR IF VM connected

Argument vm.name starts with mobileweb-

Argument host.name not equal to second-cluster-host

THEN trigger the alarm as  normal and send emails

But I'm almost sure that this is wrong, at least the host argument.

I can't figure it out how to configure the rules in order to get notified when both VMs are running on the same host.

Any help will be appreciated.

Best regards.

[scott28tt]

A cluster enabled for DRS?

If so, leave your cluster fully automated and create a VM-VM Affinity Rule: Create a VM-VM Affinity Rule

You should not then need your alarm.

[dragoscts]

Sorry but I forgot to mention that used license is a standard one so DRS is not supported.

[daphnissov]

So out of curiosity, what will you/someone do if you get such an email alert?

[dragoscts]

Get informed that both VMs are running on the same host.

Since is not critical, can be ignored or can manually migrate the VM.

I intend to create e powercli script, but the first step is this notification.

[daphnissov]

That's what I was driving at. Since you can't afford a license for DRS, which would be the "correct" way to go about this, I'd rather invest in building the automation to ensure that this is remediated *without* having to notify a person. Doing this in PowerCLI would not be complex. You could then schedule the script to run on a regular basis. Since you have vCenter Server, you also have a license to vRO and could do it there as well. And all of this without having to notify you.

[dragoscts]

I agree about the DRS but since I'm not the one who can decide this, I have to work with what is available.

I want this alarm because the script can be executed when the condition is met and a notification is send, so I know that something might be wrong.

Also, the script will contain no condition, just move vm to other node.