fix container addressability issues with cloud-init, precise, when lxc-clone is true

Bug #1431134 reported by Dimiter Naydenov
14
This bug affects 5 people
Affects Status Importance Assigned to Milestone
juju-core
Fix Released
High
Dimiter Naydenov

Bug Description

Extensive live tests on MAAS and EC2 with addressable LXC containers revealed a few issues, which need fixing before releasing 1.23-beta1:
1) bad interaction between dnsmasq from lxc (lxcbr0's DNS and DHCP server) - static IPs getting replaced after a 10.0.3.x IP lease assigned via DHCP from dnsmasq expires and gets renewed;
2) unreliable networking and/or slow startup of containers due to race conditions between cloud-init and other networking-related init jobs (especially with precise). All of those issues are exacerbated by enabling lxc-clone (enabled by default), i.e. bad UX.

After several days of testing on both EC2 and MAAS, a patch for all of those issues is about to be proposed.

Revision history for this message
Dimiter Naydenov (dimitern) wrote :
Changed in juju-core:
status: In Progress → Fix Committed
Curtis Hovey (sinzui)
Changed in juju-core:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.