What about the jrockit 4.2 testsuite run. Is this also on VMWare? How can you tell by
looking at the CC results?
I'd expect to have more test timing issues in the VMWare runs, not the other way
around.
I just saw another long run (270'), wasn't this stopped?
This is a vmware instance used only for projects that require
frequent runs,
only jboss-4.2-testsuite at this time, doesn't have parallel runs going on.
-----Original Message-----
From: Dimitris Andreadis [mailto:dandread@redhat.com]
Sent: Tuesday, June 19, 2007 2:15 PM
To: Rajesh Rajasekaran
Cc: 'JBoss.org development list'; QA
Subject: Re: Differrences in CC runs for the AS testsuite
That would explain the major time difference. What about the other machine?
Is it doing a
'single' CC run at a time or multiple ones (like different projects) ?
Rajesh Rajasekaran wrote:
> This testsuite runs currently on 2 different machines.
> Looking at the runs over the last 10 days,
> The runs which are scheduled to run frequently takes anywhere from 150 -
180
> minutes.
> The nightly run which goes on a different machine takes 220+ minutes.
> (I am disabling this nightly run to avoid confusion).
>
> At least now we have a lesser margin to worry about (150-180 min).
>
>
> -----Original Message-----
> From: jboss-qa-bounces(a)redhat.com [mailto:jboss-qa-bounces@redhat.com] On
> Behalf Of Dimitris Andreadis
> Sent: Tuesday, June 19, 2007 1:49 PM
> To:
JBoss.org development list; QA
> Subject: Differrences in CC runs for the AS testsuite
>
> The exact same CC run for jboss-4.2-testsuite-sun-1.5 with close to zero
> code changes, took
> at different times:
>
>
http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.2-testsuite-sun-1.5?l
> og=log20070619084727Lbuild.42
> 171 minutes 32 seconds
>
>
http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.2-testsuite-sun-1.5?l
> og=log20070619114948
> 159 minutes 47 seconds
>
>
http://cruisecontrol.jboss.com/cc/buildresults/jboss-4.2-testsuite-sun-1.5?l
> og=log20070619102946
> 227 minutes 51 seconds
>
> So the slowest run took 42% more time than the fastest run. Obviously this
> has an impact in
> all time sensitive tests.
>
> Is the CC machine loaded with other stuff? What else could explain the
time
> differences?
>
>