Open thatportugueseguy opened 8 months ago
You can reboot the VM in the provisioning phase, but afterwards stoping/starting needs to be done via limactl
and not just the VM itself.
I don't think there is anything that can be done about this, so I would prefer to close this issue. Thoughts @lima-vm/maintainers ?
Wonder if limactl restart
would help ?
Otherwise it would require generating the proper mount command for the new filesystems, and adopt to systemd or openrc or what have you. Might not even be possible, for some drivers. Why do we need the reboot? Muscle memory?
On a side note, systemctl reboot
?
The hostagent would have to detect the reboot, and then go through the whole "requirements" processing again once it gets an ssh connections again etc. Seems like a lot of work for something that doesn't seem to have a solid use-case.
Wonder if
limactl restart
would help ?
I've wanted this a couple of times already, even though I feel it is somewhat redundant.
Why do we need the reboot? Muscle memory?
yeah, on my part it was just bad form and muscle memory. Looking back at it, it makes sense to use reboot only on provisioning, though i'd say it's not super obvious for newcomers and might lead to situations where it's unclear what the problem is.
Is this about reverse-sshfs? virtiofs? 9p?
Is this about reverse-sshfs?
It is definitely a problem with reverse-sshfs
because the mounting is done by the guest-agent. The other 2 are probably fine.
Description
When rebooting, we lose access to the mount of the host machine home directory (
/Users/<user>
).Stopping and starting the vm again seems to fix the issue, it's probably because rebooting skips a few steps that starting and stopping does not, namely:
Minimal repro:
Using debian-11 template.