[jacorb-developer] _non_existent call hangs

[Phil Mesnier]

Hi,

If the network element is rebooted due to a failure, it is likely no TCP closures or resets will be detected by the client. So the client doesn't know the peer went away. This is nothing JacORB can address. As Nick suggested, your best bet is to set a timeout so the client will eventually abandon the request.

-Phil

> On Mar 23, 2015, at 3:35 PM, KRISHNAMURTHY, Sriram (Sriram) <sriram.krishnamurthy at alcatel-lucent.com> wrote:
> 
> Hello Nick
> 
> We are using Jacorb 3.5. 
> 
> We have network elements to which we can interface via the tmf814 interface. One of the methods that we invoke on the network Element from the Management system is getEMSSession. This is the call that internally invokes the _non_existent and hangs.
> 
> The situation was re-created by rebooting the network element. So when the Query happens to the getEMSSession and if the node reboots at the same time, there seems to be a case of this method never returning. No exceptions are seen as well. The stack trace below was from the thread dump
> 
> Please let me know if you need any more details.
> 
> Thanks
> Sriram
> 
> -----Original Message-----
> From: jacorb-developer-bounces~sriram.krishnamurthy=alcatel-lucent.com at lists.spline.inf.fu-berlin.de [mailto:jacorb-developer-bounces~sriram.krishnamurthy=alcatel-lucent.com at lists.spline.inf.fu-berlin.de] On Behalf Of Nick Cross
> Sent: Monday, March 23, 2015 12:19 PM
> To: Discussions concerning CORBA development with JacORB
> Subject: Re: [jacorb-developer] _non_existent call hangs
> 
> 
> What version are you using? Can you provide a test case?
> 
> Thanks
> 
> Nick
> 
> 
> On 23/03/15 16:03, KRISHNAMURTHY, Sriram (Sriram) wrote:
>> Hello
>> 
>> While invoking a method on the Remote Object, I see that there is a _non_existent check.
>> 
>> In cases where the remote node is abnormally terminated, I am seeing issues where the thread invoking _non_existent check doesn't return forever.
>> 
>> Is this a known issue? Is there any attribute in the jacorb.properties that would fix this issue?
>> 
>> The stack trace goes like below.
>> 
>> Thread[PollerWorker [1],10,ServerService ThreadGroup]:{MaskedEntityId ( entity ID :10.120.234.80 masked entity ID: 10.120.234.80-ASYNC )=1}
>>                  at java.lang.Object.wait(Native Method)
>>                  at org.jacorb.orb.giop.ReplyPlaceholder.getInputStream(ReplyPlaceholder.java:107)
>>                  at org.jacorb.orb.ReplyReceiver.getReply(ReplyReceiver.java:388)
>>                  at org.jacorb.orb.Delegate._invoke_internal(Delegate.java:1475)
>>                  at org.jacorb.orb.Delegate.invoke_internal(Delegate.java:1244)
>>                  at org.jacorb.orb.Delegate.invoke(Delegate.java:1232)
>>                  at org.jacorb.orb.Delegate.invokeBuiltin(Delegate.java:2275)
>>                  at org.jacorb.orb.Delegate.non_existent(Delegate.java:2223)
>>                  at org.omg.CORBA.portable.ObjectImpl._non_existent(ObjectImpl.java:61)
>>                  at com.alu.nms.server.orb.imp.mediator.CorbaTmf814GmreSessions.getEMSSession(CorbaTmf814GmreSessions.java:259)
>> 
>> 
>> 
>> Thanks
>> Sriram
>> 
>> 
>> _______________________________________________
>> jacorb-developer maillist  -  jacorb-developer at lists.spline.inf.fu-berlin.de
>> https://lists.spline.inf.fu-berlin.de/mailman/listinfo/jacorb-developer
>> 
> 
> _______________________________________________
> jacorb-developer maillist  -  jacorb-developer at lists.spline.inf.fu-berlin.de
> https://lists.spline.inf.fu-berlin.de/mailman/listinfo/jacorb-developer
> _______________________________________________
> jacorb-developer maillist  -  jacorb-developer at lists.spline.inf.fu-berlin.de
> https://lists.spline.inf.fu-berlin.de/mailman/listinfo/jacorb-developer

--
Phil Mesnier
Principal Software Engineer and Partner,   http://www.ociweb.com
Object Computing, Inc.                     +01.314.579.0066 x225