[lxc-users] Unable to copy container with snapshots to remote host

[zztest at openmailbox.org]

Hi!

I am copying several containers from one host to a remote on a LAN and 
it works well except when a container has snapshots. Local seems to 
timeout waiting for remote to finish doing something and errors out. How 
can I extend the connection timeout?

I ran:
$ lxc copy --verbose --debug wallabag02 jt3media:wallabag03
...

and ended up with:

DBUG[03-10|00:03:19] 
1.0/operations/cd8b6893-9f2b-48cf-bac2-af3c01b6dbc2/wait
error: Get 
https://jt3media:8443/1.0/operations/cd8b6893-9f2b-48cf-bac2-af3c01b6dbc2/wait: 
Unable to connect to: jt3media:8443


In addition, the remote is left in an undefined state. That is, "lxc 
list" does not show the copied container, but I cannot copy it again. I 
get an "already exists" error:

$ lxc delete jt3media:wallabag03
error: couldn't detect filesystem for 
'/var/lib/lxd/containers/wallabag03': no such file or directory

$ lxc copy wallabag02 jt3media:wallabag03
error: The container/snapshot already exists

Only the snapshot subvolume was created on remote, but deleting it does 
not clear the error.  Isn't "lxc copy" transactional? How do I reset the 
database and remove the partially complete copy?

Finally, I noticed that "migration" subvolumes are created on local for 
each copy attempt, but "lxc list" does not list them. What is the best 
way to clean those files?

I am running Ubuntu Wily with lxc v2.0.0.rc2  on btrfs on all the hosts. 
The container is about 800 MB and has one snapshot.

Thanks!
zamzing