How to get rid of false alarms on IPv6 hosts

Hello everybody.

I encountered and finally managed to solve a tricky problem. Since I did not find any description of this on the net, I'd like to share the solution here. Maybe someone else will find this useful in the future.

When I added IPv6-capable devices to a fresh Zenoss installation, they all showed up with status "host down". The hosts were in fact perfectly up and running, and I could even successfully ping them from the Zenoss web interface. But in Zenoss, they all had these big red "host down" warnings. The problem appeared on a Debian 7 machine with Zenoss installed using Hydruid's fine install script, and with the virtual appliance (both on version 4.2.5).

After digging through some python code, I found out that Zenoss relies on the host's ping6 command to return something like "rtt min/avg/max/mdev = 3.298/3.365/3.485/0.085 ms" in the last line. However it seems that the ping6 command only returns this line when the payload size (parameter -s payload) is above a certain threshold. For all Linux systems I tested (Debian 6 and 7, openSUSE 11, CentOS 5), this threshold seems to be 16 bytes, on OSX it's 8 bytes and on OpenWRT 4 bytes.

Once this is known, the solution is relatively easy: go to "Advanced->Settings->Daemons->zenping->edit config...", set the "data-length" parameter to a sufficient value (I set it to 64), save and restart the zenping process.

IMO it would make sense to set the default data-length value to something like 16 bytes or more. What do you think