Thank you for your detailed reply, I really appreciate it and I totally
understand your reasoning. Also, many thanks for a great 20.04 release
and congratulations to yourself and the team for your know-how as well
as your attitude. It's refreshing.

Just to illustrate my use case a bit further.

- I have a few VMs running as well as several LXD containers. Internal changes 
in those probably are not easily detected by zsys. The snapshot taken at boot 
is a very easy and reliable recovery point.
- They are very cheap to have since I have snapshots before and after boot as 
well. They consume no additional space. Boot up time indeed increases which 
becomes noticeable when using spinning rust (on my machines at least).

In practice I almost always rely on the snapshots taken at boot when
rolling back, but this is a laptop and it has regular boots. On my NAS I
don't rollback like ever and generally don't tinker that much.

I know about the breakage the naming convention causes with zsys, but
this was intentional :) For now at least, until I get more acquainted
with how zsys behaves, especially the gc part.

The systemd unit was a bit of a struggle and it wouldn't work unless I
added zsysd (socket error). I'll try again just in case I missed
something.

And I just try to snapshot the filesystem as early as possible. In that
sense you are absolutely correct the term "consistent state" is
debatable. It is more to the extent of "no open files".

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1875479

Title:
  Snapshot during system boot to get "known good configuration"

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/zsys/+bug/1875479/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to