Android: kill leftover webpagereplay servers.

Android: kill leftover webpagereplay servers.

Some bots downstream ended up with hundreds of instances.
Kill them all before start running the tests.

BUG=326540
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=239479

[bulach, 2013-12-06 17:41:38]

ptal

[tonyg, 2013-12-06 17:49:53]

lgtm

Looks like an excellent reliability fix!

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
File build/android/pylib/utils/test_environment.py (right):

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
build/android/pylib/utils/test_environment.py:14: for server in ['lighttpd',
'web-page-replay', 'webpagereplay']:
Do we really need both with and without the dashes? If so, maybe add a comment?

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
build/android/pylib/utils/test_environment.py:21: p.wait(1)
I didn't think wait took args. What does this do?

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
build/android/pylib/utils/test_environment.py:21: p.wait(1)
Let's parallelize this by having two loops. The first loops sends all the
signals, the second loop does all the waits.

[bulach, 2013-12-06 18:12:55]

thanks tony! all addressed, another quick look please?

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
File build/android/pylib/utils/test_environment.py (right):

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
build/android/pylib/utils/test_environment.py:14: for server in ['lighttpd',
'web-page-replay', 'webpagereplay']:
On 2013/12/06 17:49:53, tonyg wrote:
> Do we really need both with and without the dashes? If so, maybe add a
comment?

right, I think at some point it had the dashes, but that's no longer the case..
removed it.

https://codereview.chromium.org/105733013/diff/1/build/android/pylib/utils/te...
build/android/pylib/utils/test_environment.py:21: p.wait(1)
On 2013/12/06 17:49:53, tonyg wrote:
> Let's parallelize this by having two loops. The first loops sends all the
> signals, the second loop does all the waits.

good idea!
the arg is a timeout for waiting the process to finish...
I also changed and rather than 5 retries with SIGQUIT, it's now trying
increasingly more menacing signals :)

[tonyg, 2013-12-06 18:41:09]

lgtm

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
File build/android/pylib/utils/test_environment.py (right):

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
build/android/pylib/utils/test_environment.py:21: logging.warning('Killing %s
%s', server, p.pid)
I recommend we downgrade this to logging.info and also note the signal in the
message.

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
build/android/pylib/utils/test_environment.py:30: logging.warning('Failed
waiting %s %s', p.pid, e)
'Failed waiting for %s to die. %s'

[bulach, 2013-12-09 10:25:16]

thanks! comments addressed, CQing.

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
File build/android/pylib/utils/test_environment.py (right):

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
build/android/pylib/utils/test_environment.py:21: logging.warning('Killing %s
%s', server, p.pid)
On 2013/12/06 18:41:09, tonyg wrote:
> I recommend we downgrade this to logging.info and also note the signal in the
> message.

Done.

https://codereview.chromium.org/105733013/diff/40001/build/android/pylib/util...
build/android/pylib/utils/test_environment.py:30: logging.warning('Failed
waiting %s %s', p.pid, e)
On 2013/12/06 18:41:09, tonyg wrote:
> 'Failed waiting for %s to die. %s'

Done.

[commit-bot: I haz the power, 2013-12-09 10:25:45]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bulach@chromium.org/105733013/60001

[commit-bot: I haz the power, 2013-12-09 12:55:23]

Change committed as 239479