05-11-2016 02:07 PM
I am completely new to the VTM, I am testing it out to replace my netscaler VPX. I am running 2 VTM's in my AWS VPC, they are clustered and when both nodes are online they show healthy status.
I created a traffic group with a sincle elastic IP address and set one traffic manager to be active and the other to be passive.
when both traffic managers are online, everything is fine and i am able to reach the elastic IP and see the http content behind it, but when i create a failure situation and stop the instance that is holding the elastic ip, the passive traffic manager is unable to take the traffic ip address.
from the documentation i have read, this is something that should just happen and doesnt require any additional or special parameters. I am also using IAM roles instead of ec2 access/secret keys, but i dont think thats the issue because the vtm is able to see the EIPs and initiate autoscaling.
here is the error i am seeing in the traffic ip groups tab:
Error: There is a problem. Please see below for details
Error communicating with local traffic manager: There was a problem communicating with machine <Downed traffic manager>:9080: Timed out when sending request (state connection).
ERROR: Traffic IP address <EIP Address> is not hosted on any traffic managers
any help would be greately appreciated!
05-11-2016 11:38 PM
Welcome to the VTM world - I'm sure you will be quite happy with this product!
Your problem is, that you have explicit configured, that one of your clusternodes are passive - make them both active and then the VTM will automatic find the best suitable node to host your elastic IP and it will fail over in case of errors.
05-12-2016 08:23 AM
Thanks, turned out it wasn't that, it was because I did not allocate a secondary private ip address for the ENI the EIP was being attached to on the second traffic manager.
Once I allocated a secondary private ip address for the ENI i wanted the EIP to be associated with, the failover happened and took approximately 15 seconds when I stopped the instance through the ec2 console, and took about 2 seconds to failback when I brought the preferred traffic manager back online.