Slow copy/move between lxd hosts

Hi, just to ask here (dont know if this behavior is for git issue).

When i’m copying/moving containers between lxd hosts i’m getting speed only about 30MB/s
The network cards are dual port 10gbe intels in bond mode, with regular traffic between hosts i’m getting almost full speed of 10gbit.

Dell servers pretty decent.

Is there some (lxd) setting which influencing transfer speed between lxds?

I’m on freshly installed 17.10 ubuntu, all updates, lxd snap version 2.20, zfs storage backend (zfs send involved in migration), ssd disk on both servers - so imho kind of “best” configuration for lxd infrastructure.

The transfer speed was “slow” whole time i’m playing/living with lxd for (dunno) years, but right now i have some very “BIG” containers and manipulating with them is painful.


Same here with BTRFS. I think both ZFS and BTRFS have slow send/receive mechanisms. In the end, we simply create empty containers on the remote side and use rsync to copy/sync the data over. It is kinda kludgy but it works much faster than BTRFS send/receive.

If possible, can you run some ZFS send/receive commands outside of LXD to see if the problem still exists? A quick google search indicates a few threads on zfs send/receive slow. Maybe you can try to adjust the mbuffer size to see if that fixes it?


I apologise for necroposting, but maybe it may be helpful for someone in future.

The advice is simple: make sure both nodes can connect to each other, not just the one you’re transferring the container from to a new one. Once I enabled a reverse connection in the source machine’s firewall and added the source machine as a remote for the destination one, the transfer speed skyrocketed for some reason.

1 Like