Client Restart

On systems with limited resources, or that just do not want the cient to start automatically, here are the steps.

Stop client from auto-starting in both admin and user account:

sudo mv /home/user/.config/autostart/evc.desktop /home/user/.config/autostart/evc.desktop.bak && sudo mv /home/admin/.config/autostart/evc.desktop /home/admin/.config/autostart/evc.desktop.bak

Stop client from auto-starting in both admin and user account seperately:

Admin:

sudo mv /home/user/.config/autostart/evc.desktop /home/user/.config/autostart/evc.desktop.bak

User:

sudo mv /home/admin/.config/autostart/evc.desktop /home/admin/.config/autostart/evc.desktop.bak

Reversal:

sudo mv /home/user/.config/autostart/evc.desktop.bak /home/user/.config/autostart/evc.desktop && sudo mv /home/admin/.config/autostart/evc.desktop.bak /home/admin/.config/autostart/evc.desktop

There is an issue with Intel drivers on Linux systems where customers might call in regarding the client closing or restarting with any of our Linux hardware, but most likely with LC and G-Series PoE is where we’ve seen it.

The issue is documented in #d19225. You can tell that this issue is happening because you will see the following message in the syslog:

lightdm[2226]: i965: Failed to submit batchbuffer: Bad address

The message itself comes from the Mesa3D Intel drivers here:
https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/mesa/drivers/dri/i965/intel_batchbuffer.c

After the drivers print this message they immediately call exit(1) which forces us to close and there’s no way we can handle this.

The only way to prevent this is by turning off GPU decoding and VGA acceleration, but there is a performance impact to doing this. We have enabled a mitigation in #d19352 where if the client exits with an error code it will automatically start back up.

The bug for the Intel driver can be found here:
https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/1794033