RE: FW: Problem with blocking in sendReceiveTimeout

Hi Dion.

I don't think that the server is blocked - though I am not sure. The dumps I have, I get from our production environment when the system is very busy and I only have dumps from the client side.

Do you know if the mutex used by the client is a static or an instance mutex? If it is an instance mutex, it might help to have more than one server object in a 'server pool'.

 -----Original Message-----
From: 	Dion Picco [mailto:[email protected]] 
Sent:	23. december 2004 16:26
To:	Henrik Nørgaard
Cc:	[email protected]
Subject:	Re: [OB-Users] FW: Problem with blocking in sendReceiveTimeout

Hi Henrik,

On Wed, Dec 15, 2004 at 09:12:23AM +0100, Henrik Nørgaard wrote:
> If you have already received this email, I am sorry for the inconvinience. 
> I just can see my attempted posting in the archive...
> 
> Med venlig hilsen / Kind Regards, 
>  
> Henrik Hald Nørgaard
> Udvikler
>   
> Livjægergade 17A
> DK-2100 Copenhagen Ø
>  
> E-mail:  [email protected]
> Mobile: +45 2949 2838
> Phone: +45 3370 1421
> Fax: +45 7027 2760
> www.realtime.dk 
>  
> >  -----Original Message-----
> > From: 	Henrik Nørgaard  
> > Sent:	14. december 2004 12:01
> > To:	'[email protected]'
> > Subject:	Problem with blocking in sendReceiveTimeout
> > 
> > Hi Corba-experts.
> > 
> > I have a problem with a Corba-server, that tends to block as shown in 
> > the following stacktrace:
> > 
> > "Thread-38326" daemon prio=5 tid=0x02d57e58 nid=0x4d76 waiting for 
> > monitor entry [8e9d1000..8e9d19c8]
> >         at com.ooc.CORBA.GIOPClientWorkerBlocking.sendReceiveTimeout(
> > GIOPClientWorkerBlocking.java:62)
> >         - waiting to lock <0xa1b9cc40> (a com.ooc.CORBA.
> > GIOPClientWorkerBlocking)
> >         at com.ooc.CORBA.GIOPClient.request(GIOPClient.java:293)
> >         at com.ooc.CORBA.Request.invoke(Request.java:275)
> >         at com.ooc.CORBA.Delegate.non_existent(Delegate.java:231)
> >         at org.omg.CORBA.portable.ObjectImpl._non_existent(
> > ObjectImpl.java:74)
> >         at dk.realtime.duckling.backend.ServerHelper$ServerRunner.run(
> > ServerHelper.java:108)
> >         at java.lang.Thread.run(Thread.java:534)
> > 
> > I call _non_existent to check the server before calling a business-method 
> > on the Corba-object, but when the system is loaded this sometimes takes 
> > much longer than usual (several seconds) and once in a while everything 
> > seems to stop and all threads block in the above method call.
> > 
> > What is the problem? And what can be done to resolve it?
> > 
> > Med venlig hilsen / Kind Regards, 
> >  
> > Henrik Hald Nørgaard
> > Udvikler
> >   
> > Livjægergade 17A
> > DK-2100 Copenhagen Ø
> >  
> > E-mail:  [email protected]
> > Mobile: +45 2949 2838
> > Phone: +45 3370 1421
> > Fax: +45 7027 2760
> > www.realtime.dk 
> >  

What version of Orbacus are you using here?  This stack trace you are
showing here shows a client being blocked when trying to send a
request... more than likely another thread has sent a request and is
awating a reply but keeping the mutex locked.  The important thing here
is to find out if the server received the original request and if it has
sent back a reply... if your clients are blocked (as your stack trace
shows) then its possible that your servers are as well so a stack trace
here could help.  

-- 
Dion Picco, Software Engineer
IONA Technologies Inc.
Team Orbacus - Your CORBA Source
mailto:[email protected]
http://www.orbacus.com


_______________________________________________
OB-Users Mailing List - [email protected]
http://mail.ooc.nf.ca/mailman/listinfo/ob-users
Visit our support FAQ before you send a message.
http://www.orbacus.com/faq/support.html