Merge ~canonical-kernel-team/+git/autotest-client-tests:phlin/multipass-stddev into ~canonical-kernel-team/+git/autotest-client-tests:master

[Po-Hsu Lin]

BugLink: https://bugs.launchpad.net/bugs/2067978

The formula for sample standard deviation require more than 1 result,
otherwise it will cause a ZeroDivisionError.

Calculate stddev only when we have 2+ results available to avoid this issue.

[Francis Ginther (fginther)]

@cypressyew, any idea why we see the multipass test fail to start sometimes. I looked over a couple of recent runs, and we typically see a few of these fail in the beginning. I ended up filing a new bug for this: https://bugs.launchpad.net/ubuntu-kernel-tests/+bug/2068768

Given the other bug, I'm ok with this fix.

review: Approve

[Po-Hsu Lin (cypressyew)]

@Francis
I noticed that too https://bugs.launchpad.net/ubuntu-kernel-tests/+bug/2068726
From what I can see here the hardcoded Eoan might a one of the issue.

From your case it seems xenial is working, just not all of them.
We might need to dig deeper to see what's going on.

Apart from that I noticed that this test will be marked as passed with all instance were failed to launch https://bugs.launchpad.net/ubuntu-kernel-tests/+bug/2068725 maybe we should mark it as failed to avoid false impression that this test is OK.

[Po-Hsu Lin (cypressyew)]

I will land this first to reduce uncommitted changes on the server.
Patch applied and pushed.
Thanks.

diff --git a/ubuntu_performance_multipass/ubuntu_performance_multipass.py b/ubuntu_performance_multipass/ubuntu_performance_multipass.py
index b69773e..98d04a8 100644
--- a/ubuntu_performance_multipass/ubuntu_performance_multipass.py
+++ b/ubuntu_performance_multipass/ubuntu_performance_multipass.py
@@ -238,10 +238,11 @@ class ubuntu_performance_multipass(test.test):
                     minimum = min(results)
                     maximum = max(results)
                     average = sum(results) / len(results)
-                    if average > 0:
+                    if average > 0 and len(results) > 1:
                         stddev = sqrt(float(reduce(lambda x, y: x + y, map(lambda x: (x - average) ** 2, results))) / (len(results) - 1))
                         percent_stddev = (stddev / average) * 100.0 if average > 0.0 else 0.0
                     else:
+                        print("Not enough results to calculate sample standard deviation!")
                         stddev = 0.0
                         percent_stddev = 0.0
                     print("%s_%s_%s_minimum %.3f" % (test_name, release, key, minimum))