semaphoreci/flake: TestNetCustomBridge, ExpectRegex timed out after 20s

[lucab]

This is to capture a flake which is showing up recurrently on SemaphoreCI, as build logs there get periodically purged.

--- FAIL: TestNetCustomBridge (27.46s)
	rkt_net_test.go:714: Telling the child to connect via http://192.168.122.161:49152
	goroutineassistant.go:76: Command: /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/rkt --dir=/tmp/datadir-160906046 --local-config=/tmp/localdir-198078341 --system-config=/tmp/systemdir-199979808 --user-config=/tmp/userdir-596508671 --debug --insecure-options=image run --net=bridge0 --mds-register=false /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/test-tmp/rkt-inspect-networking.aci
	goroutineassistant.go:76: Command: /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/rkt --dir=/tmp/datadir-033979730 --local-config=/tmp/localdir-073695369 --system-config=/tmp/systemdir-237047892 --user-config=/tmp/userdir-100786595 --debug --insecure-options=image run --net=bridge0 --mds-register=false /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/test-tmp/rkt-inspect-networking1.aci
	goroutineassistant.go:76: Command: /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/rkt --dir=/tmp/datadir-033979730 --local-config=/tmp/localdir-073695369 --system-config=/tmp/systemdir-237047892 --user-config=/tmp/userdir-100786595 --debug --insecure-options=image run --net=bridge0 --mds-register=false /home/runner/rkt/builds/build-rkt-coreos/build-rkt-1.27.0+git/tmp/functional/test-tmp/rkt-inspect-networking2.aci
	goroutineassistant.go:64: Error encountered - shutting down
	goroutineassistant.go:65: Error: ExpectRegex timed out after 20s finding 'HTTP-Get received: (.*?)\r'.
		Output:
		
		Output: 

This started to almost-always fail on the 26th June. The last successfull PR was #3717, which merged fine and was green on master (and is unrelated to networking).

I tried to reproduce this locally, but I was unable to. And I don't see any obvious culprit for this, expect some race or some system/network configuration change on semaphore. This requires investigation directly on semaphore machines. In the meanwhile, I'm going to disable that test in order to not interfere with pending PRs.

[lucab]

TestNetCustomPtp is also reliably failing in the same way.

[squeed]

We should add some of the k8s "what is does this system have" scripts, so we dump things like the kernel version and network interfaces before test runs.

I bet there's a conflicting ip address.