Comment 5 for bug 1832694

In this case, because of suspend will fail (hang) when the graphic card switch to Nvidia, so all Nvidia related after suspend jobs were all skipped.
And then, suspend/suspend_advanced occurred when the graphic card was still Nvidia, so all after suspend jobs which depended on it, also all skipped.

I think we may switch the graphic card to another one when the graphic card failed in suspend test to avoid the kind of problem. But, it is a bit complicated.

The workaround I think now is to re-run the jobs which depends on suspend/suspend_advanced after the first run finished.