All, We have a native LXD server (3.0.0) that has the native and snap package installed. Apparently, we never ran the lxd.migrate on that server, so I'm doing some testing with a snapshot server and copying snapshots from the LXD native server (3.0.0) to an LXD native server recently installed running 3.0.3.
My ultimate goal was to copy over a few containers as snapshots. Make sure they would start and run. Stop them and then run lxd.migrate on the server. What I wasn't expecting was that when I started the snapshot on the new server, the network stack, which had contained static IP address information, now looked like a freshly installed container with the dhcp default. The containers are mostly CentOS 7 with a few Ubuntu servers. Instead of the static IP information that existed on the 3.0.0 native server, the configuration in the CentOS 7 container showed: DEVICE=eth0 BOOTPROTO=dhcp ONBOOT=yes HOSTNAME=rocketchat NM_CONTROLLED=no TYPE=Ethernet MTU= DHCP_HOSTNAME=`hostname` This information is stored in the filesystem of a CentOS 7 machine in /etc/sysconfig/network-scripts/ifcfg-eth0 My question is: Why does the static IP not follow the container, is this by design? and secondly, Is there a way to force the static IP (all network information) to be contained in the snapshot. If not, that isn't a big deal, we just have to develop procedures for taking care of this in the event that we need to take our snapshots live. All of the other services that are running in the original container are running in the snapshot and since we have a macvlan profile type in use on both servers, setting the IP to a static IP is a simple matter of repopulating the ifcfg-eth0 information. I did ultimately test the lxd.migrate on the test server which worked as expected. Thanks, Steven G. Spencer
_______________________________________________ lxc-users mailing list lxc-users@lists.linuxcontainers.org http://lists.linuxcontainers.org/listinfo/lxc-users
