Print logs to debug shutdown integration test

[2opremio]

to track down #1889.

[2opremio]

And of course, now the test passes. Sigh ...

Will rebuilt one more time and just merge if it passes again.

[2opremio]

It failed on the second attempt: https://circleci.com/gh/weaveworks/scope/5357?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link

>>> Test ./110_shutdown_test.sh finished after 60.1 secs with error: exit status 1
Check scope exits cleanly within 10 seconds
2546fa33c2dd9c04da13e0a49457ab445141411785ec8f0ca5c940ca31da4115
Weave Scope is reachable at the following URL(s):
  * http://10.240.0.6:4040/
<probe> INFO: 2016/09/23 16:18:23.923461 command line args: --logtostderr=true --mode=probe --probe.docker=true
<probe> INFO: 2016/09/23 16:18:23.923672 probe starting, version 7fddaa9, ID 340ff198d57020ff
<probe> INFO: 2016/09/23 16:18:23.923716 publishing to: localhost:4040
<probe> ERRO: 2016/09/23 16:18:23.951093 Error fetching app details: Get http://127.0.0.1:4040/api: dial tcp 127.0.0.1:4040: getsockopt: connection refused
<probe> WARN: 2016/09/23 16:18:23.985040 Cannot resolve scope.weave.local.: dial tcp 172.17.0.1:53: getsockopt: connection refused
<probe> WARN: 2016/09/23 16:18:23.985305 Error collecting weave status, backing off 10s: Get http://127.0.0.1:6784/report: dial tcp 127.0.0.1:6784: getsockopt: connection refused
<probe> INFO: 2016/09/23 16:18:24.137258 Success collecting weave ps
INFO: 2016/09/23 16:18:24.339891 === received SIGINT/SIGTERM ===
*** exiting
<probe> INFO: 2016/09/23 16:18:24.340508 probe exiting
test #1 "docker_on host1-5357-0.us-central1-a.scope-integration-tests logs weavescope 2>&1 | grep 'app exiting' || (docker_on host1-5357-0.us-central1-a.scope-integration-tests logs weavescope 2>&3 ; false)" failed:
    program terminated with code 1 instead of 0
1 of 3 tests failed in 59.903s.

It seems like the app didn't start at all?

[rade]

isn't this just a race between

scope_on $HOST1 launch
scope_on $HOST1 stop

in the test?

Launching is async since runs docker run -d ..., so stop can come along before runsvinit has started one or both of the processes.

A short sleep would 'cure' this.

FWIW, in weave net, launch waits until a status http requests on the container succeeds, or the container stops.

[2opremio]

isn't this just a race between

scope_on $HOST1 launch
scope_on $HOST1 stop

Good point, I will add the sleep and try to reproduce.

[2opremio]

And as usual @rade was right and the sleep fixed the problem (three runs without failing).

I want to leave the debug output for the future.

@rade PTAL

[rade]

^