mms,mss supported on java 1.6

[John Franey]

Is there a known problem of running mss, mms, jboss 5 on java 1.6?

Is there a web page to show how to work around?

I'm trying the media-servlet-jsr309 example in java.16 environment on linux.  I get a Segmentation Fault.  If this is not a known problem, I can dig into it.  If its a known problem, whats the workaround?

Thanks,

john

[Sachin Parnami]

Please print stack trace to find out exact problem :)

--
Regards,
Sachin Parnami

[John Franey]

On Tue, Dec 15, 2009 at 12:35 AM, Sachin Parnami <sachin...@gmail.com> wrote:

Please print stack trace to find out exact problem :)

Sachin,

Thanks.  Of course.  I'll take your response to mean that mobicents mms, mss running in jboss 5 on jvm 1.6 is a configuration that is known to work.  Knowing that this configuration is supported on jvm 1.6 gives me confidence that my effort to get a stack trace will not be wasted.

Getting the stack trace is a problem, I think.  Any ideas on how to do so?  Is there a jvm flag that will cause jvm to print a stack trace when it suffers a segmentation fault?   If you have done this, can you say how?  or provide a link to a web-site?  Be assured I will google for the right configuration. 

Regards,

John

[aayush bhatnagar]

Take a heap dump. If your app crashed a *.hprof file must have been
created with the details of the fault.

Another thing...which OS are you using ?

--
aayush
http://in.linkedin.com/in/abhatnagar19
======

[John Franey]

Sachin,

Currently, until I find the core dump, or learn how to get jvm to dump core, this is all of the diagnostic that is current available to me: the standard out logging of jboss (server.log doesn't have greater detail):

from="127.0.1.1:5060" 

to="127.0.1.1:5080" 

time="1260880543055"

isSender="false" 

transactionId="z9hg4bk1584431560" 

callId="835584564" 

firstLine="INVITE sip:10...@127.0.1.1:5080 SIP/2.0" 

>

<![CDATA[INVITE sip:10...@127.0.1.1:5080 SIP/2.0

Via: SIP/2.0/UDP 192.168.2.102:5060;rport=5060;branch=z9hG4bK1584431560;received=127.0.1.1

From: <sip:jo...@192.168.2.102>;tag=1626707527

To: <sip:10...@127.0.1.1:5080>

Call-ID: 835584564

CSeq: 20 INVITE

Contact: <sip:jo...@192.168.2.102:5060>

Content-Type: application/sdp

Max-Forwards: 70

User-Agent: Linphone/3.1.2 (eXosip2/3.3.0)

Subject: Phone call

Content-Length: 290

v=0

o=john 123456 654321 IN IP4 192.168.2.102

s=A conversation

c=IN IP4 192.168.2.102

t=0 0

m=audio 7078 RTP/AVP 111 110 0 8 101

a=rtpmap:111 speex/16000/1

a=rtpmap:110 speex/8000/1

a=rtpmap:0 PCMU/8000/1

a=rtpmap:8 PCMA/8000/1

a=rtpmap:101 telephone-event/8000

a=fmtp:101 0-11

]]>

</message>

07:35:43,237 INFO  [InitialRequestDispatcher] Dispatching the request event to org.mobicents.servlet.sip.example.MediaJSR309Application

07:35:43,273 INFO  [PlayerServlet] MediaPlaybackServlet: Got request:

INVITE

07:35:43,291 INFO  [Mobicents-SIP-Servlets] <message

from="127.0.1.1:5080" 

to="127.0.1.1:5060" 

time="1260880543290"

isSender="true" 

transactionId="z9hg4bk1584431560" 

callId="835584564" 

firstLine="SIP/2.0 180 Ringing" 

>

<![CDATA[SIP/2.0 180 Ringing

Via: SIP/2.0/UDP 192.168.2.102:5060;rport=5060;branch=z9hG4bK1584431560;received=127.0.1.1

From: <sip:jo...@192.168.2.102>;tag=1626707527

To: <sip:10...@127.0.1.1:5080>;tag=64288601_0f2c25a3_83943032417409

Call-ID: 835584564

CSeq: 20 INVITE

Contact: <sip:127.0.1.1:5080;transport=udp>

Content-Length: 0

]]>

</message>

07:35:43,307 INFO  [InitialRequestDispatcher] Request event dispatched to org.mobicents.servlet.sip.example.MediaJSR309Application

Segmentation fault

In my working directory there is no core dump file.  The linux man page lists a number of conditions when the core file is not created, but none of these I find to be true (for example, jboss process has write permission in the working directory).  

Regards,

John

On Tue, Dec 15, 2009 at 12:35 AM, Sachin Parnami <sachin...@gmail.com> wrote:

[John Franey]

On Tue, Dec 15, 2009 at 7:48 AM, aayush bhatnagar <abhatnag...@gmail.com> wrote:

Take a heap dump. If your app crashed a *.hprof file must have been
created with the details of the fault.

There is no *.hprof file in current working directory after the crash.

I'm looking into how to create a heap dump under this condition.  I don't expect to succeed because the jvm is suffering a segmentation fault which is an unstable condition of the process.  I expect linux to dump a core file, but it isn't.  I was hoping there were someone reading this list that knows how to go about it (perhaps avoiding a fruitless effort on my part).

 

Another thing...which OS are you using ?

 

uname -a:

Linux isidore 2.6.31-16-generic #52-Ubuntu SMP Thu Dec 3 22:07:16 UTC 2009 x86_64 GNU/Linux

java -version:

java version "1.6.0_14"

Java(TM) SE Runtime Environment (build 1.6.0_14-b08)

Java HotSpot(TM) 64-Bit Server VM (build 14.0-b16, mixed mode)

[aayush bhatnagar]

There is no hprof created in the /bin directory of jboss?

[John Franey]

On Tue, Dec 15, 2009 at 8:28 AM, aayush bhatnagar <abhatnag...@gmail.com> wrote:

There is no hprof created in the /bin directory of jboss?

On 15/12/2009, John Franey <jjfr...@gmail.com> wrote:
> On Tue, Dec 15, 2009 at 7:48 AM, aayush bhatnagar <
> abhatnag...@gmail.com> wrote:
>
>> Take a heap dump. If your app crashed a *.hprof file must have been
>> created with the details of the fault.
>>
>>
> There is no *.hprof file in current working directory after the crash.

Correct.

[Jean Deruelle]

worth trying on JDK 1.6.0_17 but I don't expect it to be better 

myself I use 

Linux deruelle-laptop 2.6.31-16-generic-pae #53-Ubuntu SMP Tue Dec 8 05:20:21 UTC 2009 i686 GNU/Linux

and 

java version "1.6.0_10"

Java(TM) SE Runtime Environment (build 1.6.0_10-b33)

Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode)

and it is working fine w/o any segmentation fault

[Vladimir Ralev]

Actually there is a specific version of JBoss AS to be used with jdk6.
I think right now we use the 1.5 one. Although I dont think this is
the issue here it is worth a try.
http://sourceforge.net/projects/jboss/files/JBoss/JBoss-5.1.0.GA/ the
"-jdk6" one.

[John Franey]

On Tue, Dec 15, 2009 at 4:11 PM, Vladimir Ralev <vladimi...@gmail.com> wrote:

Actually there is a specific version of JBoss AS to be used with jdk6.
I think right now we use the 1.5 one. Although I dont think this is
the issue here it is worth a try.
http://sourceforge.net/projects/jboss/files/JBoss/JBoss-5.1.0.GA/ the
"-jdk6" one.

Maybe. 

Is there a script in mobicents source that builds/copies the mobicents components into a clean jboss?

I used eclipse debugger to zero down into the call that causes the segfault.  not the same as a stack trace, but perhaps useful.

Not quite sure where to go from here (probably to try the jdk6 jboss), but for the record, here is stack information.  The JVM segfaults while handling a call to DatagramSocket.send invok at this point:

Daemon Thread [JainMgcpStackImpl-FixedThreadPool-thread-1] (Suspended (breakpoint at line 336 in JainMgcpStackImpl))

JainMgcpStackImpl.send(DatagramPacket) line: 336

CreateConnectionHandler(TransactionHandler).send(JainMgcpCommandEvent) line: 485

TransactionHandler.access$1100(TransactionHandler, JainMgcpCommandEvent) line: 76

TransactionHandler$ScheduleCommandSend.perform() line: 786

CreateConnectionHandler(TransactionHandler).run() line: 400

ThreadPoolExecutor$Worker.runTask(Runnable) line: 886

ThreadPoolExecutor$Worker.run() line: 908

Thread.run() line: 619

This is the datagram packet being passed in:

packet DatagramPacket (id=232) address Inet4Address (id=234)

                (/127.0.1.1) buf (id=238) bufLength 396 length 396 offset 0 port 2427

Buffer content is:

CRCX 1 /mobicents/media/packetrelay/$@127.0.1.1:2427 MGCP 1.0

C: 1

M: sendrecv

N: isi...@127.0.1.1:2727

v=0

o=john 123456 654321 IN IP4 192.168.2.102

s=A conversation

c=IN IP4 192.168.2.102

t=0 0

m=audio 7078 RTP/AVP 111 110 0 8 101

a=rtpmap:111 speex/16000/1

a=rtpmap:110 speex/8000/1

a=rtpmap:0 PCMU/8000/1

a=rtpmap:8 PCMA/8000/1

a=rtpmap:101 telephone-event/8000

a=fmtp:101 0-11

 

On Tue, Dec 15, 2009 at 3:36 PM, Jean Deruelle <jean.d...@gmail.com> wrote:

[John Franey]

On Tue, Dec 15, 2009 at 8:36 AM, Jean Deruelle <jean.d...@gmail.com> wrote:

worth trying on JDK 1.6.0_17 but I don't expect it to be better 

myself I use 

Linux deruelle-laptop 2.6.31-16-generic-pae #53-Ubuntu SMP Tue Dec 8 05:20:21 UTC 2009 i686 GNU/Linux

and 

java version "1.6.0_10"

Java(TM) SE Runtime Environment (build 1.6.0_10-b33)

Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode)

and it is working fine w/o any segmentation fault

All on Ubuntu 9.10 latest update (as of 10/19):

Linux isidore 2.6.31-16-generic #53-Ubuntu SMP Tue Dec 8 04:02:15 UTC 2009 x86_64 GNU/Linux

I checked out:  http://mobicents.googlecode.com/svn/tags/servers/media/2.x.y/jsr-309/mobicents-jsr309-impl-2.0.0.BETA2

Ran 'mvn test' with the following jvm:

No segmentation fault on 1.6.0_13 (but a test failure occurred):

java version "1.6.0_13"

Java(TM) SE Runtime Environment (build 1.6.0_13-b03)

Java HotSpot(TM) 64-Bit Server VM (build 11.3-b02, mixed mode)

Segmentation fault with 1.6.0_14:

java version "1.6.0_14"

Java(TM) SE Runtime Environment (build 1.6.0_14-b08)

Java HotSpot(TM) 64-Bit Server VM (build 14.0-b16, mixed mode)

Segmentation fault with 1.6.0_17:

java version "1.6.0_17"

Java(TM) SE Runtime Environment (build 1.6.0_17-b04)

Java HotSpot(TM) 64-Bit Server VM (build 14.3-b01, mixed mode)

Obviously, a jvm regression.  I've submitted a bug at sun citing this mobicents test case; I will post if they accept the bug report. 

Also, I was able to get a core dump.  I learned that ubuntu disables core dump by default.  To enable: ulimit -S -c 100000000

Regards,

John

[Bartosz Baranowski]

Great :) Thanks for help with mms. Could You post TCK failure details ? if its valid failure there is chance it will make it to B3 (fixed).

--
Bartosz Baranowski
JBoss R & D
==================================
Word of criticism meant to improve is always step forward.

[John Franey]

On Sat, Dec 19, 2009 at 5:30 PM, Bartosz Baranowski <bara...@gmail.com> wrote:

Great :) Thanks for help with mms. Could You post TCK failure details ? if its valid failure there is chance it will make it to B3 (fixed).

 

I'm thinking its not a valid error.  It occurs 2 out of 5 runs.  Maybe a timing issue, right?

-------------------------------------------------------------------------------

Test set: org.mobicents.javax.media.mscontrol.MsControlFactoryImplTest

-------------------------------------------------------------------------------

Tests run: 6, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 0.108 sec <<< FAILURE!

testMediaConfigClone(org.mobicents.javax.media.mscontrol.MsControlFactoryImplTest)  Time elapsed: 0.005 sec  <<< ERROR!

java.lang.RuntimeException: Failed to find a local port to bound stack

        at org.mobicents.mgcp.stack.JainMgcpStackImpl.init(JainMgcpStackImpl.java:157)

        at org.mobicents.mgcp.stack.JainMgcpStackImpl.createProvider(JainMgcpStackImpl.java:284)

        at org.mobicents.jsr309.mgcp.MgcpStackFactory.getMgcpStackProvider(MgcpStackFactory.java:81)

        at org.mobicents.javax.media.mscontrol.MsControlFactoryImpl.<init>(MsControlFactoryImpl.java:63)

        at org.mobicents.javax.media.mscontrol.spi.DriverImpl.getFactory(DriverImpl.java:30)

        at org.mobicents.javax.media.mscontrol.MsControlFactoryImplTest.setUp(MsControlFactoryImplTest.java:34)

        at junit.framework.TestCase.runBare(TestCase.java:128)

        at junit.framework.TestResult$1.protect(TestResult.java:110)

        at junit.framework.TestResult.runProtected(TestResult.java:128)

        at junit.framework.TestResult.run(TestResult.java:113)

        at junit.framework.TestCase.run(TestCase.java:120)

        at junit.framework.TestSuite.runTest(TestSuite.java:228)

        at junit.framework.TestSuite.run(TestSuite.java:223)

        at org.junit.internal.runners.OldTestClassRunner.run(OldTestClassRunner.java:35)

        at org.apache.maven.surefire.junit4.JUnit4TestSet.execute(JUnit4TestSet.java:62)

        at org.apache.maven.surefire.suite.AbstractDirectoryTestSuite.executeTestSet(AbstractDirectoryTestSuite.java:140)

        at org.apache.maven.surefire.suite.AbstractDirectoryTestSuite.execute(AbstractDirectoryTestSuite.java:127)

        at org.apache.maven.surefire.Surefire.run(Surefire.java:177)

        at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)

        at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)

        at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)

        at java.lang.reflect.Method.invoke(Method.java:597)

        at org.apache.maven.surefire.booter.SurefireBooter.runSuitesInProcess(SurefireBooter.java:345)

        at org.apache.maven.surefire.booter.SurefireBooter.main(SurefireBooter.java:1009)

[John Franey]

Here is the url to the bug report:  http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6912878

It was accepted only a few days ago and posted into the database today.

Please vote for this bug.

Regards,

John

[Jean Deruelle]

Voted up

[John]

I've tested for this problem on mms 2.0.0 BETA3, and jdk 1.6.0_18
(both released over the last week or so).

Unfortunately, the segmentation fault occurs on the above combination,
too.

java version "1.6.0_18"
Java(TM) SE Runtime Environment (build 1.6.0_18-b07)
Java HotSpot(TM) 64-Bit Server VM (build 16.0-b13, mixed mode)

Linux isidore 2.6.31-16-generic #53-Ubuntu SMP Tue Dec 8 04:02:15 UTC
2009 x86_64 GNU/Linux

Regards,
John

On Dec 26 2009, 8:39 am, Jean Deruelle <jean.derue...@gmail.com>
wrote:
> Voted up
>
>
>
> On Thu, Dec 24, 2009 at 4:43 PM, John Franey <jjfra...@gmail.com> wrote:
>
> > On Sat, Dec 19, 2009 at 5:07 PM, John Franey <jjfra...@gmail.com> wrote:

>
> >> On Tue, Dec 15, 2009 at 8:36 AM, Jean Deruelle <jean.derue...@gmail.com>wrote:
>
> >>> worth trying on JDK 1.6.0_17 but I don't expect it to be better
> >>> myself I use
> >>> Linux deruelle-laptop 2.6.31-16-generic-pae #53-Ubuntu SMP Tue Dec 8
> >>> 05:20:21 UTC 2009 i686 GNU/Linux
> >>> and
> >>> java version "1.6.0_10"
> >>> Java(TM) SE Runtime Environment (build 1.6.0_10-b33)
> >>> Java HotSpot(TM) Server VM (build 11.0-b15, mixed mode)
>
> >>> and it is working fine w/o any segmentation fault
>
> >> All on Ubuntu 9.10 latest update (as of 10/19):
>
> >> Linux isidore 2.6.31-16-generic #53-Ubuntu SMP Tue Dec 8 04:02:15 UTC 2009
> >> x86_64 GNU/Linux
>
> >> I checked out:

> >>http://mobicents.googlecode.com/svn/tags/servers/media/2.x.y/jsr-309/...

[Amit Bhayani]

Voted
I faced this error with JAIN SIP (SIP RA) Stack too

[Vladimir Ralev]

Is it possible for you to use 32-bit Linux and JVM?

[John Franey]

On Sun, Jan 17, 2010 at 11:19 AM, Vladimir Ralev <vladimi...@gmail.com> wrote:

Is it possible for you to use 32-bit Linux and JVM?

Interesting question.  Not sure what you mean.

If you mean, is it possible for me to test to see if 32-bit linux and jvm suffers the same problem, no.  I do not have that resource at this time.  I'll leave it to someone who has that platform handy.

If you mean, is it possible for me to run a 32-bit linux and jvm in production, maybe.

Regards,

John

[Amit Bhayani]

This works fine. I mean I tried 32 bit Ubuntu on 64bit machine and 32 bit JVM.