I had mentioned an issue that I've run into at the end of an unrelated question a few weeks ago but I think it's appropriate to make a new discussion about it. I thought that this problem had disappeared (or I was mis-reading what I was seeing) so I didn't follow up, but my QA guys ran into it again so apparently it hasn't. The simple description of the problem is that VMs that have been configured at installation with static IP addresses are periodically reverting to DHCP addresses without any apparently change in configuration. On first boot I'm running "/opt/vmware/share/vami/vami_ovf_process --setnetwork" and that seems to configure the IP on the primary interface just fine. But after runnng for a few days the appliances will suddenly (literally, I've had this happen when I was actually SSHed into them and tailing logs) change IP to something in the subnet's DHCP scope. When this happens, if I log into VAMI on its new IP and look at the network tab, this is what I see:
Note that 192.168.64.247 is NOT the configured static IP and IS in our DHCP scope (and I see a proper lease for it in the DHCP server). The contents of the /etc/network/interfaces file (this is an Ubuntu 10.04 appliance) is:
If I reboot the appliance it comes back up immediately on the static IP 192.168.64.161. What the heck might VAMI be doing that would cause this? It's definitely saying that the IP is not managed by VAMI but I don't know of any other services on the appliance that would override, on the fly, the static configuration of eth0. This is a total show-stopper for us (assuming at least some of our customers will have DHCP set up where they probably shouldn't or will be installing these as POCs in a lab with DHCP) and we have customers clamoring for these appliances in the very near future. Anyone have any ideas what I should be checking? Nothing I've looked at in any logs that I can find show what might be doing it. Thanks everyone!