geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sandip Ghayal <sgha...@yahoo.com>
Subject RE: [jira] Commented: (GERONIMO-553) Security Timeout Test failures
Date Fri, 28 Jan 2005 11:32:07 GMT
Well I did some debugging and its slowness of my
machine. 

My machine is not able to accurately time 300ms from
the time 1 second timeout was set.

I also found during investigation that sometimes even
the timer expires at 1300-1400ms. (I think that scan
timer expired late)

I don't thing if having such an accurate timeout is a
requirement. and I don't 6.86 secs is too much time.
(other tests in this group also take around same or
even more time to run)

Thanks,

Sandip
--- "Alan D. Cabrera" <adc@toolazydogs.com> wrote:

> Nope.  The point is to see if the security code
> removes the security
> context all by itself.  Using wait/notify would
> require wiring test code
> directly into the server.
> 
> The real question is why it takes 1.2s for code that
> should take
> milliseconds.
> 
> 
> Regards,
> Alan
> 
> > -----Original Message-----
> > From: Dain Sundstrom
> [mailto:dsundstrom@gluecode.com]
> > Sent: Thursday, January 27, 2005 8:04 PM
> > To: dev@geronimo.apache.org
> > Subject: Re: [jira] Commented: (GERONIMO-553)
> Security Timeout Test
> > failures
> > 
> > Can this test be rewritten using wait/notify or a
> phantom reference?
> > 
> > -dain
> > 
> > --
> > Dain Sundstrom
> > Chief Architect
> > Gluecode Software
> > 310.536.8355, ext. 26
> > 
> > On Jan 27, 2005, at 9:24 AM, Sandip Ghayal wrote:
> > 
> > > Well we are not doing anything much
> > >
> > > Just sleeping for 3 second and then sleeping for
> 7
> > > seconds.
> > >
> > > Ok I have experimented with new values.
> > >
> > > Scan time of 100 ms
> > > Timeout value: 2000 ms
> > > Sleep 1: 1000 ms
> > > Sleep 2: 3000 ms
> > >
> > > Total test time  6.86 sec
> > >
> > > I think this should be acceptable.
> > >
> > > Sorry Alan, don't want any one screaming at you
> :-)
> > >
> > > Cheers,
> > >
> > > Sandip
> > > --- Jeremy Boynes <jboynes@apache.org> wrote:
> > >
> > >> I have to ask - what are we doing that means we
> need
> > >> to make the
> > >> timeouts so long? Is there some intrinsically
> long
> > >> running operation
> > >> here or are things just inefficient?
> > >>
> > >> I ask 'cos a 15 second test is going to get
> Alan
> > >> yelled at again :-)
> > >>
> > >> --
> > >> Jeremy
> > >>
> > >>
> > >> Sandip Ghayal wrote:
> > >>> Hi Alan,
> > >>>
> > >>> Increasing timeout scan did not make any
> > >> difference.
> > >>> It is one second timeout value that is causing
> > >> test to
> > >>> fail.(though I do agree timeout scan of 50ms
> will
> > >> be
> > >>> hard on slow computers, which we can increase)
> > >>>
> > >>> I do feel that timeouts should be in terms of
> > >> couple
> > >>> of seconds and not just 1 second.
> > >>>
> > >>> We can surely optimize its value even further,
> but
> > >> I
> > >>> think 5 second should be good enough for test.
> it
> > >>> takes close to 15secs on my machine to run
> this
> > >> test
> > >>> (with 5 sec timeout)
> > >>>
> > >>> Cheers,
> > >>>
> > >>> Sandip
> > >>> --- "Alan Cabrera (JIRA)"
> > >> <dev@geronimo.apache.org>
> > >>> wrote:
> > >>>
> > >>>
> > >>>>     [
> > >>>>
> > >>>
> > >>>
> > >>
> > >
> http://issues.apache.org/jira/browse/GERONIMO-553?
> > > page=comments#action_58168
> > >>>
> > >>>> ]
> > >>>>
> > >>>> Alan Cabrera commented on GERONIMO-553:
> > >>>> ---------------------------------------
> > >>>>
> > >>>> 50ms timeout scan?  Yikes, that means that
> slower
> > >>>> computers will be doing nothing but scanning.
>  Can
> > >>>> you revert the parameters for the test back
> and
> > >> set
> > >>>> the timeout scan to 100ms and see if it
> passes?
> > >>>>
> > >>>>
> > >>>>> Security Timeout Test failures
> > >>>>> ------------------------------
> > >>>>>
> > >>>>>         Key: GERONIMO-553
> > >>>>>         URL:
> > >>>>
> > >>>>
> http://issues.apache.org/jira/browse/GERONIMO-553
> > >>>>
> > >>>>>     Project: Apache Geronimo
> > >>>>>        Type: Improvement
> > >>>>>  Components: security
> > >>>>> Environment: Windows XP
> > >>>>>    Reporter: Sandip Ghayal
> > >>>>>    Priority: Minor
> > >>>>> Attachments: security_Timout_fix.patch,
> > >>>>
> > >>>> security_Timout_fix.patch
> > >>>>
> > >>>>> When during Geronimo Build I found the
> Security
> > >>>>
> > >>>> test failures for Timeout Test.
> > >>>>
> > >>>>> During investigation what I found was my
> login
> > >>>>
> > >>>> context was being dropped during running of
> the
> > >>>> test. When investigating further I found that
> the
> > >>>> timeout was set to one second. Where as the
> next
> > >>>> request came after 1.2 secs (though the sleep
> in
> > >> the
> > >>>> test was for 300 ms).
> > >>>>
> > >>>>> I think that the time out values and machine
> > >> speed
> > >>>>
> > >>>> is too critical for this test to work with
> such
> > >>>> precisions. So I propose to increase the
> timeout
> > >>>> value to 10 secs and increase the sleep times
> > >> within
> > >>>> the test to 3 secs and 17 secs. (just
> multiplied
> > >> all
> > >>>> times by 10).
> > >>>>
> > >>>> --
> > >>>> This message is automatically generated by
> JIRA.
> > >>>> -
> > >>>> If you think it was sent incorrectly contact
> one
> > >> of
> > >>>> the administrators:
> > >>>>
> > >>>>
> > >>>
> > >>>
> > >>
> 
=== message truncated ===



		
__________________________________ 
Do you Yahoo!? 
Yahoo! Mail - Helps protect you from nasty viruses. 
http://promotions.yahoo.com/new_mail

Mime
View raw message