-
Notifications
You must be signed in to change notification settings - Fork 56
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
manifest: wait for user input before rebooting when creating a bootc installer #714
Conversation
I'm not sure of the exact setup where this is expected to be used, but in my experience you do want it to reboot when it is finished installing. In those cases you set things up so that the installer iso is booted once and on reboot it uses the disk that was just installed. |
The issue could also be solved by putting the choice on the user: osbuild/bootc-image-builder#438 We could expose the |
Silly question, since this will wait for user input, is there a message that will say "free to remove removable media" or something like that, so that you can easily remove the USB stick / press continue? I'm assuming you would remove whatever you have plugged in, hit next and now you have your OS. |
|
Pulling the plug without a proper shutdown isn't safe. Anaconda keeps the target disk mounted until it shuts down or reboots. |
This PR is stale because it has been open 30 days with no activity. Remove "Stale" label or comment or this will be closed in 7 days. |
What's the decision here? Do we still want this? |
In issue osbuild#230 it was raised that it is undesirable to automatically reboot when the install is finished. So instead halt the install and show a message to the user. Closes: osbuild/bootc-image-builder#230
Out automatic tests need a way to know when the install is finished. Before we could just trigger on a qmp:REBOOT event but with this change the installer will wait for user input before doing a reboot. Ejecting the cdrom tray is also done after the user confirmation so triggering on qmp:DEVICE_TRAY_MOVED is also not an option. Go the conservative route and just output on the serial port. This has the added benefit that users on the serial port also get an indication when the install is done.
This PR is stale because it has been open 30 days with no activity. Remove "Stale" label or comment or this will be closed in 7 days. |
This commit adds a new wait `event:serial-output` event to the QEMU vm class and uses it in the `test_iso_installs` test to accommodate the change to the ISO installer [0] that it no longer automatically reboots but waits for the user to interact with the system. [0] osbuild/images#714
This commit adds a new wait `event:serial-output` event to the QEMU vm class and uses it in the `test_iso_installs` test to accommodate the change to the ISO installer [0] that it no longer automatically reboots but waits for the user to interact with the system. [0] osbuild/images#714
This commit adds a new wait `event:serial-output` event to the QEMU vm class and uses it in the `test_iso_installs` test to accommodate the change to the ISO installer [0] that it no longer automatically reboots but waits for the user to interact with the system. [0] osbuild/images#714
This commit adds a new wait `event:serial-output` event to the QEMU vm class and uses it in the `test_iso_installs` test to accommodate the change to the ISO installer [0] that it no longer automatically reboots but waits for the user to interact with the system. [0] osbuild/images#714
Small followup for osbuild#714 - to add an extra level of safety call "sync" before signaling that the install is finished.
This commit adds a new wait `event:serial-output` event to the QEMU vm class and uses it in the `test_iso_installs` test to accommodate the change to the ISO installer [0] that it no longer automatically reboots but waits for the user to interact with the system. [0] osbuild/images#714
In issue #230 it was raised that it is undesirable to automatically reboot when the install is finished. So instead halt the install and show a message to the user.
This is a draft because it changes the existing behavior so users who rely on the auto-reboot (like our tests) will fail [0] - but if everyone is happy and we consider the old behavior a bug I will undraft it.
Closes: osbuild/bootc-image-builder#230
[0] We can fix this via https://github.com/osbuild/bootc-image-builder/compare/main...mvo5:handle-halt-instead-of-reboot?expand=1