Resurrecting a corrupted global database

Hi there,

I’m using LXD 3.2 on Ubuntu 16.04 LTS installed via snap. When I was trying to create a new container with:

lxc launch ubuntu:18.04 test

LXD hit a out of disk condition. I’m using a simple “dir” storage-pool on ext4. Afterwards LXD doesn’t start anymore. Moving the global database into a different place fixed this, but all my containers are gone.

How can I revive the database or start from scratch and re-create my containers from filesystem?

If you still have your global database files as they were just after the out of disk conditions, please make a tarball of them and attach them here (or email them to me - free.ekanayaka at canonical dot com ). I’ll try to see if the can be recovered.

I would like to, but I’m a little bit unsure, if I can share the content. What’s inside this database? Is there any documentation about the content/structure?

The database contains metadata about your containers (name, configuration, etc). It does not contain information about what’s inside the container (rootfs data etc).

Does this include the network setup, profiles and how are containers connected?

Yes, it includes the networks and profiles you have defined.

I’m sorry, I cannot share the global database. But I’m willing to debug the issue. Is there anything I can do?