Snap auto refresh to lxd 3.17 breaks production cluster, cannot create new containers

After auto refresh to lxd 3.17 automatically while creating container getting following error:

 $ lxc launch ubuntu:18.04 test1
Creating test1
Error: Failed container creation: Create container: 
Create LXC container: UNIQUE constraint failed: 
storage_volumes.storage_pool_id, storage_volumes.node_id, 
storage_volumes.project_id, storage_volumes.name, 
storage_volumes.type

After  removing this entry from cluster:
$ lxd sql global "delete from storage_volumes where name='test1';"

Running again the lxc container launch fails with an error.

$lxc launch ubuntu:18.04 test1
Creating test1
Error: Failed container creation: Create container: 
Create LXC container: 
Failed to add device 'eth0': No such object

I have create an issue #6512 on github under lxd project. I believe many people might be facing the same problem. Also based on my experience this is the 3rd time cluster has broken due to snap auto refresh of lxd. My suggestion will be to disable it by default instead of enable it. Also suggest in the document not to enable it for production environments.

This issue is resolved by removing the manually created bridge and dnsmasq entries in snap directory as it is a cheap way used by lxd daemon to determine if the network is managed or unmanaged. As my network is self-managed (i.e. not managed by lxd), it was giving an error due to presence of those directory and files.

Just for people upgrading from 3.16 to 3.17 be careful if you are using internal directory inside the snap folder of lxd, better not to use it.