checkbox-remote is not able to reconnect after rebooting when prime-select to intel on output display via nvidia

[jeremyszu]

My environment is a workstation.
The DP output is through nvidia card.

---

$ lspci -nnv -d ::0x0300
00:02.0 VGA compatible controller [0300]: Intel Corporation Device [8086:4c8a] (rev 04) (prog-if 00 [VGA controller])
 DeviceName: Onboard IGD
 Subsystem: Hewlett-Packard Company Device [103c:88be]
 Flags: bus master, fast devsel, latency 0, IRQ 142
 Memory at 6012000000 (64-bit, non-prefetchable) [size=16M]
 Memory at 4000000000 (64-bit, prefetchable) [size=256M]
 I/O ports at 4000 [size=64]
 Capabilities: <access denied>
 Kernel driver in use: i915
 Kernel modules: i915

01:00.0 VGA compatible controller [0300]: NVIDIA Corporation Device [10de:2206] (rev a1) (prog-if 00 [VGA controller])
 Subsystem: Hewlett-Packard Company Device [103c:88d7]
 Flags: bus master, fast devsel, latency 0, IRQ 255
 Memory at 9d000000 (32-bit, non-prefetchable) [size=16M]
 Memory at 6000000000 (64-bit, prefetchable) [size=256M]
 Memory at 6010000000 (64-bit, prefetchable) [size=32M]
 I/O ports at 3000 [size=128]
 Expansion ROM at 9e080000 [disabled] [size=512K]
 Capabilities: <access denied>
 Kernel modules: nvidiafb, nouveau, nvidia_drm, nvidia

---

Checkbox job will prime-select to intel, then the nvidia-prime will blacklist nvidia driver.
```
$ cat /lib/modprobe.d/blacklist-nvidia.conf
# Do not modify
# This file was generated by nvidia-prime
blacklist nvidia
blacklist nvidia-drm
blacklist nvidia-modeset
alias nvidia off
alias nvidia-drm off
alias nvidia-modeset off
```

Therefore, gdm will not ablt to get the drm device.
```
/usr/lib/gdm3/gdm-x-session[1563]: (EE) [drm] Failed to open DRM device for pci:0000:01:00.0: -19
```

My assumption is:
checkbox-ng.service try to launch a GUI application for fwts, but when prime-select intel. There is no display available.
Thus, the checkbox-ng.service is dead.
```
$ systemctl status checkbox-ng.service
● checkbox-ng.service - Checkbox remote slave service
     Loaded: loaded (/lib/systemd/system/checkbox-ng.service; disabled; vendor preset: enabled)
     Active: inactive (dead)
```

The result is checkbox-remote is waiting for client back
```
---------------------------[ Running job 108 / 187 ]----------------------------
----------[ Switch GPU to Intel Corporation PCI ID 0x4c8a and reboot ]----------
...
13:06:48 DEBUG Executing: 'sleep 5; reboot'...
13:06:48 Connection lost!
13:07:01 INFO:master:master: Connection lost due to: connection closed by peer
13:07:01 connection closed by peer
...

16:44:43 INFO:master:master: Connection lost due to: [Errno 111] Connection refused
16:44:43 Reconnecting...
16:44:43 INFO:master:master: Connection lost due to: [Errno 111] Connection refused
```

In this case, all GTK function should not works.

[Maksim Beliaev (beliaev-maksim)]

Bug was migrated to GitHub: https://github.com/canonical/checkbox/issues/178.
Bug is no more monitored here.