This question is somewhat related to a separate question I revived today.
I’ve been using LXC (without LXD) on Fedora Linux for years. I keep things simple. All of my containers use simple ext4 filesystem directories, like this:
/var/lib/lxc/vps00/rootfs/... /var/lib/lxc/vps01/rootfs/... /var/lib/lxc/vps02/rootfs/... ... vps = virtual private server
And whenever I need to, I’ll clone a vpsXX instance using the humble
cp -rp command (with the container off, of course); then update the destination’s MAC and IP-Address.
Although I prefer remaining with my straight LXC strategy (i.e. using the
/usr/bin/lxc-xxx CLI approach), I’m looking into LXD today because I upgraded my Bare-Metal O/S from Fedora-30 to Fedora-32, and now my vpsXX LXC’s won’t start. This happens every once in a while.
It’s likely a conflict with the LXC
config file and something that changed with Fedora-32, though I can’t seem to figure out the issue.
Now I don’t need fancy image cloning or snapshotting (it’s all on the same SSD H/W RAID-0 Stripe anyway). Given that I like keeping things simple, transparent and not over-engineered (which I get with LXC), are there things that LXD would offer me that I should consider? For example, performance when several containers are running simultaneously? (Just an example).
PS: I may have no choice if I can’t figure out how to get my LXC containers running again on Fedora-32 (which were running perfectly fine right before the upgrade from Fedora-30). Any help with ideas on that problem would be welcome, too.
Thank you in advance for you inputs! Sorry for the long question.