Why is ZXTM pinging my back-ends when I don't have a PING health monitor attached?

There are in fact two ways that ZXTM can produce ICMP Echos.

One way to generate pings to your back-end(s)is to apply a ping health monitor, that periodically pings the nodes of the pool to which it is applied, in order to determine whether these nodes are alive.

However, the other element of ZXTM which generates pings is the health messages. Whilst the actual health messages are relayed between ZXTMs in a cluster via UDP multicast, the ZXTMs themselves must check their connectivity to the outside network and the back-end servers - and in the current release of ZXTM this is done by pinging them. The process is that if a ZXTM cannot ping its Default Gateway then it will assume that it has failed and drop any Traffic IP addresses running on it (which will then be raised by another ZXTM in the cluster), and if a ZXTM cannot ping all of its back-ends then the same will happen.

As soon as ZXTM receives a response to a ping from a back-end, it will pass the back-end test, so every backend isn't pinged on every health check.

ZXTM allows you to configure a different IP address to ping instead of the %gateway%, but it will not allow you to divert pings from the back-end servers.

Can ZXTM raise its Traffic IPs if it cannot perform this health check?

The answer to this question is, no. Upon failure to raise a traffic IP, ZXTM will output this message:

ERROR: Traffic IP Address <IP address> is not hosted on any Traffic Managers.

Please make sure that ZXTM is able to send an ICMP to both your front-end gateway and your back-end nodes before attempting to raise a traffic IP.

Chris Buckley [Zeus Support] 24 July 2006