Odd, looks like it's been passing for several runs since then. I don't think this is flakiness because those tests just look for a .ko file. Instead, I assume this has to do with the kernel 4.14.21 merge that Guenter was trying to get through in those runs since those autotests check for available kernel modules, and lakitu uses that kernel version.

(for the record the CL is https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/934882)

Agree with #2 - this is a real failure. I marked the CL a V-1 (Looks like I can't remove the CQ+1).
Note that this is a showstopper for kernel merges, and thus for incoming security fixes. It also increases the risk for future merges since we'll deviate more and more from mainline until the problem is fixed. Increasing priority accordingly.

assign to groeck@ to close the bug when the bad CL is fixed and merged.

nxia@: It is worth noting that my CL is the victim here, not the culprit.

Also see b/73888655.

Ok, I'll track the progress on b/73888655.