You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-dev@axis.apache.org by Andreas Leitner <no...@sbox.tugraz.at> on 2002/08/16 15:19:02 UTC
Re: Session Lifecycle vs. ServiceLifecycle ?
Hi Matthias,
hi axis-dev
Note: I am CCing this to axis-dev.
Matthias and I found that in axis beta3 the function "ServiceLifecycle.destroy" seems not to be called at all. Here is a the transcript of our conversation from the axis user list. I think it includes all details necessary to reproduce the problem.
My questions is
Are we doing something wrong? If not, is this a known (and possibly fixed in cvs) bug?
many thanks in advance and keep up the great work,
Andreas
On Fri, Aug 16, 2002 at 01:54:35PM +0200, Matthias Brunner wrote:
> Hello Andreas,
>
> On Friday 16 August 2002 12:20, Andreas Leitner wrote:
> >
> > Again thanks. I am able to trigger "init"s correctly. No problem here. But
> > I have not every succeeded triggering "destroy" other than calling it
> > directly. Is there any way to reproduce a termination of the session?
> >
> > What I have done so far:
> >
> > *) Produced a Axis Web Service with cookie session scope
> > *) Set the timeout of the session (via Session.setTimeout) in "init" to 1
> > (that function is called - I can see it in the log) *) then from a dot net
> > client (which generates his proxy via the autogenerated wsdl file) I
> > initiate a connection and call some function. on the first call of some
> > dummy SOAP function, "init" shows up in the log. then i close the app and
> > wait... wait... wait (and "destroy" does not show up in the log, and thus
> > is not beeing called) *) status quo: I have no idea how I should go to
> > trigger a call to "destroy"
> >
>
> I have just attached a debugger to tomcat and set a breakpoint in the
> destroy() method of my service. destroy() seems never to be called. (although
> the tomcat manager webapp tells me that the session is no longer active)
> It is quite possible that this is a bug, someone :-) should post to the dev
> list or submit a bug report.
> Maybe it is already fixed in the CVS.
>
> Cheers!
> --
> Matthias Brunner <mb...@blumenstrasse.vol.at>
> PGP FP 7862 32B3 3B75 292A F76F 5042 8587 21AB 5B89 D501
> Check out http://blumenstrasse.vol.at/~mb/gpgkey.asc
>
>
Re: Session Lifecycle vs. ServiceLifecycle ?
Posted by Matthias Brunner <mb...@blumenstrasse.vol.at>.
Steve,
On Monday 19 August 2002 18:07, Steve Loughran wrote:
> 1. can you verify this against CVS head.
I verified it against the current nightly build (2002-08-19). Have there been
any changes since then?
> 2. who can think of nice automatable way to verify this in a junit test
> case? I suppose a side effect like creating a file might be a viable
> technique.
>
I can think of using a static variable which is incremented when init() is
called and decremented when destroy() is called, with a very low timeout
limit set. It has to be 0 at the end of the test.
Cheers,
Matthias
>
> ----- Original Message -----
> From: "Matthias Brunner" <mb...@blumenstrasse.vol.at>
>
> On Friday 16 August 2002 15:19, Andreas Leitner wrote:
> >
> > Matthias and I found that in axis beta3 the function
> > "ServiceLifecycle.destroy" seems not to be called at all. Here is a the
> > transcript of our conversation from the axis user list. I think it
>
> I verified this for current CVS and submitted a bug report:
> http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11815
--
Matthias Brunner <mb...@blumenstrasse.vol.at>
PGP FP 7862 32B3 3B75 292A F76F 5042 8587 21AB 5B89 D501
Check out http://blumenstrasse.vol.at/~mb/gpgkey.asc
Re: Session Lifecycle vs. ServiceLifecycle ?
Posted by Steve Loughran <st...@iseran.com>.
Mathias
1. can you verify this against CVS head.
2. who can think of nice automatable way to verify this in a junit test
case? I suppose a side effect like creating a file might be a viable
technique.
validating shutdown code is always hard; debugging it equally dubious. AFAIK
the destroy stuff is called when the AxisServlet is destroyed, which is
something the container does. But there could be scope for some extra
session lifecycle management, maybe with a plugin that works with servlet2.3
alongside 2.2
----- Original Message -----
From: "Matthias Brunner" <mb...@blumenstrasse.vol.at>
To: <ax...@xml.apache.org>
Sent: Monday, August 19, 2002 5:30 AM
Subject: Re: Session Lifecycle vs. ServiceLifecycle ?
On Friday 16 August 2002 15:19, Andreas Leitner wrote:
> Hi Matthias,
> hi axis-dev
>
> Note: I am CCing this to axis-dev.
>
> Matthias and I found that in axis beta3 the function
> "ServiceLifecycle.destroy" seems not to be called at all. Here is a the
> transcript of our conversation from the axis user list. I think it
includes
> all details necessary to reproduce the problem.
>
> My questions is
>
> Are we doing something wrong? If not, is this a known (and possibly fixed
> in cvs) bug?
>
I verified this for current CVS and submitted a bug report:
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11815
--
Matthias Brunner <mb...@blumenstrasse.vol.at>
PGP FP 7862 32B3 3B75 292A F76F 5042 8587 21AB 5B89 D501
Check out http://blumenstrasse.vol.at/~mb/gpgkey.asc
Re: Session Lifecycle vs. ServiceLifecycle ?
Posted by Matthias Brunner <mb...@blumenstrasse.vol.at>.
On Friday 16 August 2002 15:19, Andreas Leitner wrote:
> Hi Matthias,
> hi axis-dev
>
> Note: I am CCing this to axis-dev.
>
> Matthias and I found that in axis beta3 the function
> "ServiceLifecycle.destroy" seems not to be called at all. Here is a the
> transcript of our conversation from the axis user list. I think it includes
> all details necessary to reproduce the problem.
>
> My questions is
>
> Are we doing something wrong? If not, is this a known (and possibly fixed
> in cvs) bug?
>
I verified this for current CVS and submitted a bug report:
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=11815
--
Matthias Brunner <mb...@blumenstrasse.vol.at>
PGP FP 7862 32B3 3B75 292A F76F 5042 8587 21AB 5B89 D501
Check out http://blumenstrasse.vol.at/~mb/gpgkey.asc