RE: [JBoss-user] Multihome JBoss Issue

2003-12-07 Thread Rod Macpherson
Running multiple instances and netstat shows a 1098 and 1099 listener on each 
host so that looks good. One caveat is that running the same configuration twice 
generates complains about the DefaultDS. That is reasonable but it should be noted. So 
the only outstanding problem is the use of JNDI properties to create InitialContext 
objects. In the multihome configuration you have to use the no-arg InitialContext. 
 
No bind errors but datasource whines:
 
run -c foo --host bar
run -c bar --host foo 
 
No problem:
 
run -c foo --host bar
run -c baz --host foo 
 
No problem, three distinct listeners show up for 1099 and 1098
 
run -c foo --host bar
run -c baz --host foo 

run -c default --host sandbox
 
 
 
-Original Message- 
From: Scott M Stark [mailto:[EMAIL PROTECTED] 
Sent: Sun 12/7/2003 11:14 AM 
To: [EMAIL PROTECTED] 
Cc: 
Subject: Re: [JBoss-user] Multihome JBoss Issue



Which says the rmi stub is bound to 1.1.1.21 (could be 1.2.11.21 and
the mail is munging the address?). If this is not the sandbox host
address, then it needs to be overriden using the java.rmi.server.hostname
system property. This need to be looked into further as the --host
value should be setting the bind address of the rmi stub as well. What
does netstat -an | egrep '(1099|1098)' show on this box?


Scott Stark
Chief Technology Officer
JBoss Group, LLC


Rod Macpherson wrote:

> Here's the gook.
>
> C:\ Telnet sandbox 1099
>
> 
¼fsr?java.rmi.MarshalledObject|+?ùfcn>??I?hashlocBytest?[BobjBytesq~?xpë?°Tur?[B¼=?°Ta?xp)¼fthttp://sandbox:8083/q~q
> ~uq~?+¼fsr 
org.jnp.server.NamingServer_Stub??xrjava.rmi.server.RemoteStubT¦_+ïße?xr?java.rmi.server.RemoteObject+a¦æ
>  
   a3??
> xpw3
> UnicastRef1.1.1.21?J?¢v0·?üOÇx
>
> Connection to host lost.



---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


<>

RE: [JBoss-user] Multihome JBoss Issue

2003-12-07 Thread Rod Macpherson
I changed the IP addresses frequently when testing this and might have reported the 
wrong one in the previous email. Anyay, we are running JBoss on XP and W2K and same 
issue appears on both. We narrowed this problem down to where it appears when using 
JNDI properties in a new InitialContext. Supplying the host name form the system 
variable (jboss.bind.address) fails as does hard-coding that hostname. A "connection 
refused" appears as soon as you try to lookup a home interface on a session bean, for 
example. 
 
Creating the default InitialContext solved the problem as Bill Burke suggested. We did 
not want to assume the web container and the EJB container were in the same JVM which 
is why we were using the JNDI properties in the first place -- not everywhere of 
couse, just on things that could be remote.
 
I just setup an IP on my home system mapped to an alternate IP address. In this case 9 
is the new address so this looks correct. Problem is that if I create an 
InitialContext and pass in propreties that specify 9 (that is to say the hostname 
bound to that in my local hosts file) I will get connection refused. 
 
 
C:\>netstat -an | grep 1099
  TCP68.140.254.9:1099  0.0.0.0:0  LISTENING
C:\>netstat -an | grep 1098
  TCP68.140.254.9:1098  0.0.0.0:0  LISTENING
 
 

-Original Message- 
From: Scott M Stark [mailto:[EMAIL PROTECTED] 
Sent: Sun 12/7/2003 11:14 AM 
To: [EMAIL PROTECTED] 
Cc: 
Subject: Re: [JBoss-user] Multihome JBoss Issue



Which says the rmi stub is bound to 1.1.1.21 (could be 1.2.11.21 and
the mail is munging the address?). If this is not the sandbox host
address, then it needs to be overriden using the java.rmi.server.hostname
system property. This need to be looked into further as the --host
value should be setting the bind address of the rmi stub as well. What
does netstat -an | egrep '(1099|1098)' show on this box?


Scott Stark
Chief Technology Officer
JBoss Group, LLC


Rod Macpherson wrote:

> Here's the gook.
>
> C:\ Telnet sandbox 1099
>
> 
¼fsr?java.rmi.MarshalledObject|+?ùfcn>??I?hashlocBytest?[BobjBytesq~?xpë?°Tur?[B¼=?°Ta?xp)¼fthttp://sandbox:8083/q~q
> ~uq~?+¼fsr 
org.jnp.server.NamingServer_Stub??xrjava.rmi.server.RemoteStubT¦_+ïße?xr?java.rmi.server.RemoteObject+a¦æ
>  
   a3??
> xpw3
> UnicastRef1.1.1.21?J?¢v0·?üOÇx
>
> Connection to host lost.



---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


<>

Re: [JBoss-user] Unexpected Signal : 11

2003-12-07 Thread Ricardo Argüello
Brian Wallis wrote:

Does this always happen on the same system or on more than one. You can see 
these sort of errors if you have a memory or other hardware problem. One good 
test for memory related issues is to compile a linux kernel, that usually 
gets a sig11 if you have any marginal memory in the machine.
Or you could try Memtest86 if you have an x86 machine:
http://www.memtest86.com/
Ricardo Argüello



---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


Re: [JBoss-user] Unexpected Signal : 11

2003-12-07 Thread Brian Wallis
On Friday 05 December 2003 06:09, [EMAIL PROTECTED] wrote:
> Brian, this problem occurs only when it is running under Linux. When we
> changed our glib from glibc-2.3.2-4.80.6 to glibc-2.3.2-4.80.8 at RedHat 8
> using up2date -u the problem got worse.
>
> I looked at
> http://developer.java.sun.com/servlet/SessionServlet?url=/developer/bugPara
>de/bugs/4927116.html
>
> But nothing worked, they say:
> -"Release Fixed  1.4.2_04, tiger-beta"
>
> And:
> "we don't see this behavior with J2SE 1.4.0 or 1.4.1_05"
>
> I tryed all this options, but 1.4.2_04 doesn't exists, and using 1.4.0 or
> 1.4.1_05 it crashes too.

Sounds like it is not a related problem. Do you get the same error message 
mentioned in that bug report? 

  We are running JBOSS 3.2.1.  When we have duplicate class, we get the 
  following error message
  java.lang.LinkageError: duplicate class definition:
com/nortel/mftP/dmlP/modelP/faultP/workflowP/AlarmClearWorkflowC

> Any ideas?

Not really. I'm running Mandrake 9.2 with glibc 2.3.2-14mdk so we are on much 
the same glibc version (I doubt there is much difference) and I have not seen 
a sig 11 for quite a while and have been running CHO and load tests on the 
system. What kernel version are you running. I'm using the redhat 7.2 kernel 
which is 2.4.7 (Rational's clearase requires this kernel).

Does this always happen on the same system or on more than one. You can see 
these sort of errors if you have a memory or other hardware problem. One good 
test for memory related issues is to compile a linux kernel, that usually 
gets a sig11 if you have any marginal memory in the machine.

brian...



---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


[JBoss-user] Running a RMI server into JBoss

2003-12-07 Thread Marco Tedone
Hi, is there a best practice to run a RMI server into Jboss? I'm thinking of
realizing a RMI application to manager an email server, and would like to
run it into Jboss as a service. Any suggestion as what shall I do?

Thanks,

Marco





---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


Re: [JBoss-user] Container Responsibility for CMR Foreign Keys

2003-12-07 Thread Alexey Loubyansky
Unfortunately, you can't use this kind of key generation with 
insert-after-ejb-post-create.
[ 784322 ] INSERT after ejbPostCreate
http://sourceforge.net/tracker/?func=detail&aid=784322&group_id=22866&atid=381174

By the spec (and current implementation), the primary key should be set 
in ejbCreate and available in ejbPostCreate. Which is impossible in the 
case there was no INSERT after ejbCreate.
The entity key generation commands that work with 
insert-after-ejb-post-create are pk-sql and key-generator.

Hanson, Matthew wrote:
Hi,

Per your request, I have tested a bean with auto-increment primary key
fields and without foreign-key mappings.
I created a mini-project for Jboss 3.2.2 comprised of only a single bean
class that should have Hypersonic auto-increment and create a primary key
for the bean.  The bean has no relationships at all and does have the
auto-increment tag.  Additionally, just for thoroughness, all of the
 tags in standardjboss.xml are set to true.
The null primary key exception was thrown again.


---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user


Re: [JBoss-user] Multihome JBoss Issue

2003-12-07 Thread Scott M Stark
Which says the rmi stub is bound to 1.1.1.21 (could be 1.2.11.21 and
the mail is munging the address?). If this is not the sandbox host
address, then it needs to be overriden using the java.rmi.server.hostname
system property. This need to be looked into further as the --host
value should be setting the bind address of the rmi stub as well. What
does netstat -an | egrep '(1099|1098)' show on this box?

Scott Stark
Chief Technology Officer
JBoss Group, LLC

Rod Macpherson wrote:

Here's the gook. 

C:\ Telnet sandbox 1099

ÂÏsrâjava.rmi.MarshalledObject|ââÃÏcâ>ââIâhashlocBytestâ[BobjBytesq~âxpÃ?ÂÎurâ[BÂââÂTÎâxp)ÂÏthttp://sandbox:8083/q~q
~uq~ââÂÏsr 
org.jnp.server.NamingServer_Stubââxrjava.rmi.server.RemoteStubÎâââÃÃeâxrâjava.rmi.server.RemoteObjectâaâÃ
 
   a3ââ
xpw3
UnicastRef1.1.1.21âJâÂâ0ââÃÎÃx
Connection to host lost.


---
This SF.net email is sponsored by: IBM Linux Tutorials.
Become an expert in LINUX or just sharpen your skills.  Sign up for IBM's
Free Linux Tutorials.  Learn everything from the bash shell to sys admin.
Click now! http://ads.osdn.com/?ad_id=1278&alloc_id=3371&op=click
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user