[pve-devel] [RFC PATCH qemu-server 0/4] add reboot api

[Dominik Csapak]

this adds a reboot api call for vms, which uses a reboot trigger file
that gets detected by the qm cleanup call to start the vm again

this api call is useful when users want to apply pending hardware changes
without waiting for the vm to shutdown

i send this as rfc because i am not sure with the following things:
(probably makes more sense to read this after looking at the patches)
* how to properly handle locks in this case?
  afaics with containers, we do not, as in there could be a race
  (albeit a short one) where the user could start/delete/move the
  container when rebooting it
  here i also simply trust that the code runs fast enough for the user
  not be able to do anything in between...
  we could of course also start with '--skiplock', while still holding
  the lock during cleanup
* makes the whole design sense? should we instead start our own task instead
  and poll the shutdown/start procedure?

Dominik Csapak (4):
  qm: cleanup: detect and handle reboot trigger
  api: add reboot api call
  qm: add reboot command
  api: add missing index child links

 PVE/API2/Qemu.pm | 49 ++++++++++++++++++++++++++++++++++++++++++++++++
 PVE/CLI/qm.pm    | 15 +++++++++++++++
 2 files changed, 64 insertions(+)

-- 
2.20.1