Catch-22 problem with health monitor
We are trying to overcome a "catch-22" situation with a load balancer VIP - any suggestions will be appreciated...
The VIP has a pool with 2 member nodes. The health monitor tests a specific URL on each node so that if one goes down all traffic will be sent to the other node.
The problem is when BOTH nodes are down (for example when starting the application after maintenance) - the application startup has a dependency on a call back to the load balancer VIP to retrieve the same URL that is used by the health monitor. But the health monitors haven't realized the URL is back up so traffic doesn't get to the member nodes and the application start-up fails.
So we need to use the health monitors to support fail-over if one of the nodes goes down during normal operation, but if both nodes go down we need to keep sending traffic through to allow the application to start back up.
Thanks in advance...