Additional Experiments:
- No NAT, lxc bridges to VM eth0, which is bridged to host machine br0
Still fails
- Let timeout on unregister_notifier keep going
Still running after an hour, can't create new containers (original issue)
sudo conntrack -{L,F} hangs...
- Purge conntrack flowtable on host right before unmount
Still fails
- Look at conntrack table on host when its stuck
Had issues with this because of hang
Next steps:
- Track down TCP connection that's being held open
Additional Experiments:
- No NAT, lxc bridges to VM eth0, which is bridged to host machine br0
Still fails
- Let timeout on unregister_notifier keep going
Still running after an hour, can't create new containers (original issue)
sudo conntrack -{L,F} hangs...
- Purge conntrack flowtable on host right before unmount
Still fails
- Look at conntrack table on host when its stuck
Had issues with this because of hang
Next steps:
- Track down TCP connection that's being held open