<qyliss>
kinda amused at how, if cloud-hypervisor fails to boot a VM, it'll just try rebooting it a few times to see if that works
<alj[m]>
Well better than just giving up after encountering a slight hick-up right?
<qyliss>
i'd usually rather programs just crash, so they can be restarted by something with more context
<qyliss>
for example, cloud-hypervisor here requires another process to be running first that terminates after one connection, so when it tries to reboot, it'll fail because that process has gone away
<qyliss>
whereas if it just crashed, then a supervisor (if present) could restart both in the right order
<qyliss>
but mostly I was amused just because it's very unusual behaviour
<IdleBot_cef0f9d0>
CRIU requires careful figuring out if what you do fits well in the CRIU limitations; might be limiting the amount of brave souls trying to use it?
<IdleBot_cef0f9d0>
And I believe they do not even support X — although maybe with xmove one can solve this — and Wayland should be harder…
<pie_>
most exmaples seem to use xvnc
<IdleBot_cef0f9d0>
Pretty natural choice! When I played with running OpenOffice on a eInk reader, I also used Xvnc…
<IdleBot_cef0f9d0>
Limited environments differ, but the natural solutions are often general