> So I wonder is this "just" a conflict between how libvirt expects pools to be set up (and as it does by itself) vs the manual set up one?
(1) If libvirt is only supposed to work with a top-level pool, then it should have refused to allow me to create a libvirt pool with a slash in the zfs pool name. But that's not a great solution. Given that everything apart from zpool usage stats works fine, it would be much better to allow use of a parent dataset, which can be done by fixing the zpool command invocation. This is much more flexible: it allows the zpool to be shared with other applications, and have separate quotas and usage reporting for libvirt versus those other applications. I definitely wouldn't want to dedicate an entire zpool to libvirt. (2) In any case, if an error occurs when libvirt is refreshing a storage pool and shelling out to zfs subcommands, the failure should be reported and propagated. Ideally it wouldn't discard the previous volume info either. Any system error which can happen, will sooner or later happen :-) -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1767997 Title: virt-manager destroys all volumes in libvirt zfs pool To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1767997/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs