I'm starting to use CoreOS (on AWS ECS). After having it launch my first container, I see this in journalctl
:
Could not generate persistent MAC address for vethXXXX: No such file or directory
Here's more context. I've removed the time and instance information, but this is all within the same second. Note there are two distinct veth
entries. I don't know if that means anything.
systemd[1]: Started docker container 1234
systemd[1]: Starting docker container 1234
dockerd[595]: time="2015-07-23T23:30:52Z" level=info msg="GET /v1.17/containers/1234/json"
dockerd[595]: time="2015-07-23T23:30:52Z" level=info msg="+job container_inspect(1234)"
systemd-timesyncd[473]: Network configuration changed, trying to establish connection.systemd-udevd[7501]: Could not generate persistent MAC address for vethYYYY: No such file or directory
kernel: device vethXXXX entered promiscuous mode
kernel: IPv6: ADDRCONF(NETDEV_UP): vethXXXX: link is not ready
systemd-udevd[7508]: Could not generate persistent MAC address for vethXXXX: No such file or directory
systemd-networkd[497]: vethXXXX: Configured
kernel: eth0: renamed from vethYYYY
kernel: IPv6: ADDRCONF(NETDEV_CHANGE): vethXXXX: link becomes ready
kernel: docker0: port 2(vethXXXX) entered forwarding state
kernel: docker0: port 2(vethXXXX) entered forwarding state
systemd-networkd[497]: vethXXXX: Gained carrier
I found a discussion of this error on Ubuntu and it comes down to removing a udev rule, which doesn't seem to exist on CoreOS. There's a discussion about iptables with OpenVPN, which again doesn't seem to apply. Here's a bridge rule for LXC on Ubuntu. Again, I don't see how to apply that.
I haven't done anything with the networkd or flannel configuration. If the problems are in that area, I need specific steps on how to fix it for use in AWS ECS.
lxd
– Yancey