On Fri, Oct 23, 2009 at 2:16 PM, Andrew Scherkus <[email protected]>wrote:

> On Fri, Oct 23, 2009 at 12:28 PM, Nicolas Sylvain 
> <[email protected]>wrote:
>
>>
>>
>> On Fri, Oct 23, 2009 at 12:21 PM, Andrew Scherkus 
>> <[email protected]>wrote:
>>
>>> I've never witnessed these tests taking an extra 10-20 seconds on my
>>> local machine, no.
>>>
>>> I don't doubt that some of the tests might be flaky themselves, but that
>>> machine does run tests slower.  Take a look at the SVG tests, for example:
>>>
>>> http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html#tests=LayoutTests%2Fsvg
>>>
>>> Are there other tricks I do on my local machine to simulate running on
>>> the bots?  I usually try to test for these things by maxing out my CPU then
>>> running layout tests but even then they run smoothly.
>>>
>>
>> This machine is one of the oldest linux machine we have in the lab. I'll
>> recreate it, make sure it run fast, and see if it helps.
>>
>
> Thanks Nicolas.
>

I changed the machine, and so far so good. The time to run the layout tests
on this machine went from 11minutes to 5 minutes. Looks like
something was wrong on the machine after all.  There are still some
flakiness... but at least they are not TIMEOUT.

Thanks

Nicolas


>
> I'm wondering if it's slow disk access... know of any simple commands I can
> use to simulate disk thrashing?
>
>
>>
>> Nicolas
>>
>>
>>> Andrew
>>>
>>> On Fri, Oct 23, 2009 at 12:02 PM, Nicolas Sylvain <[email protected]
>>> > wrote:
>>>
>>>>
>>>>
>>>> On Fri, Oct 23, 2009 at 11:59 AM, Andrew Scherkus <
>>>> [email protected]> wrote:
>>>>
>>>>> I've been trying to get the media layout tests passing consistently,
>>>>> but WebKit Linux (dbg)(3) takes an absurdly longer time to run tests and I
>>>>> don't know why.
>>>>>
>>>>> For example:
>>>>>
>>>>> http://src.chromium.org/viewvc/chrome/trunk/src/webkit/tools/layout_tests/flakiness_dashboard.html#tests=video-played
>>>>>
>>>>> To keep the tree green (and collect data), I've marked all media layout
>>>>> tests on Linux debug as pass/fail/timeout.  My hope is if the bot was less
>>>>> bogged down, it would lead to faster build times (GTTF) and less flaky
>>>>> results/timeouts (LTTF).
>>>>>
>>>>> This machine is supposed to be fast.
>>>>
>>>> Are you saying that this flakiness never happens on your machine?
>>>>
>>>> Are you sure the bot is really to blame here?
>>>>
>>>> Nicolas
>>>>
>>>> Any ideas?
>>>>> Andrew
>>>>>
>>>>>
>>>>> >>>>>
>>>>>
>>>>
>>>
>>
>

--~--~---------~--~----~------------~-------~--~----~
Chromium Developers mailing list: [email protected] 
View archives, change email options, or unsubscribe: 
    http://groups.google.com/group/chromium-dev
-~----------~----~----~----~------~----~------~--~---

Reply via email to