Breaks machine without IPv4: "Route info failed"
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
cloud-init |
New
|
Undecided
|
Unassigned |
Bug Description
When I try to deploy a MAAS node on a pure IPv6 network, it seems to install normally and reboot, but cloud-init gives off an error about route_info failing. Eventually the console moves on to a login prompt, but the machine is not reachable on the network — which means I can't log in at all.
The console shows:
«
Cloud-init v. 0.7.5 running 'init-local' at Fri, 03 Oct 2014 04:19:29 <etc>
cloud-init-
* Starting Mount network filesystems
* Stopping Mount network filesystems
cloud-init-
* Starting configure network device
Cloud-init v. 0.7.5 running 'init' at Fri, 03 Oct 2014 04:19:30 <etc>
ci-info: +++++++
ci-info: +------
ci-info: | Device | Up | Address | Mask | Hw-Address |
ci-info: +------
ci-info: | lo | True | 127.0.0.1 | 255.0.0.0 | |
ci-info: | eth0 | True | - | - | 00:12:34:56:78:90 |
ci-info: +------
ci-info: !!!!!!!
»
After that it pauses for a long time, and finally moves on to a login prompt.
Here's what was installed in the node's /etc/network/
«
auto lo
auto eth0
iface eth0 inet6 static
netmask 64
address fd0d:1777:
gateway fd0d:1777:
»
The node doesn't show up even in neighbour discovery. If I do the same thing but with eth0 also configured to get a dynamic IPv4 address from DHCP, then it boots up normally and becomes reachable through both IPv4 and IPv6.
Related branches
- Graham Binns (community): Approve
-
Diff: 107 lines (+40/-15)2 files modifiedsrc/provisioningserver/drivers/osystem/debian_networking.py (+21/-5)
src/provisioningserver/drivers/osystem/tests/test_debian_networking.py (+19/-10)
description: | updated |
Addendum: the cloud-config URL was based on a hostname (which resolved in both IPv4 and IPv6), not an IPv6 address. I'll try again but without IPv4 resolution for that hostname.