[Mulgara-general] Can't connect to remote Mulgara repository using ConnectionFactory

[Paul Gearon]

Disabling the firewall is one approach. What about pushing everything
onto the same port, like Ronald suggested?
  --rmiobjectport 1099

Regards,
Paul

On Wed, Nov 19, 2008 at 8:15 PM, David Legg
<david.legg at searchevent.co.uk> wrote:
> Hi Ron,
>
> Thanks for the quick response and the WireShark suggestion!
>
> I've installed WireShark and fully disabled the Windows firewall
> (instead of just enabling port 1099 like last time).  I'm still getting
> the same exception but WireShark shows some sort of two way RMI
> interaction going on between the client and server: -
>
>  RMI  JRMI, ReturnData
>  TCP  hotu-chat > rmi registry [ACK]
>  RMI  JRMI, Ping
>  RMI  JRMI, PingAck
>  RMI  JRMI, DgcAck
>  RMI JRMI, Ping
>  TCP rmi registry > hotu-chat [ACK]
>  RMI  JRMI, PingAck
>  RMI  JRMI, Call
>  RMI  JRMI, ReturnData
>  RMI  JRMI, Ping
>  RMI  JRMI, PingAck
>  RMI  JRMI, DcgAck
>  TCP  rmi registry > hotu-chat [ACK]
>  TCP  hotu-chat > rmi registry [RST, ACK]
>
> That looks like the two machines are talking to each other OK although
> that last entry is coloured Red which indicates a Reset.
>
> Regards,
> David Legg
>
>
> Life is hard, and then you die wrote:
>> On Thu, Nov 20, 2008 at 12:32:54AM +0000, David Legg wrote:
>>
>>> I'm having trouble getting a test program to talk to a remote Mulgara
>>> repository and wondered if anyone has any suggestions.
>>>
>>> The test program is based on the one mentioned on the Wiki [1] and
>>> explicitly located here [2]
>>>
>>> This is my first attempt to use the ConnectionFactory and it looks like
>>> it is falling over somewhere in the call to
>>> org.mulgara.connection.ConnectionFactory.newConnection().  I'm using the
>>> latest revision from SVN (revision 1378) on the server side and an
>>> identical set of jar files on the client side.
>>>
>>> Thinking that perhaps I've got the wrong URL I've tried all sorts of
>>> combinations from rmi://optimus:1099/server1 to
>>> rmi://192.168.1.10/server1 but it makes little difference to the
>>> result.  I've disabled firewalls and I've even run telnet to see if I
>>> could access the remote port (which I could).  I've also run netstat -l
>>> on the remote (Ubuntu) server and it lists something listening on port
>>> 1099 (well, actually it says :::1099 which makes me think it is an IPv6
>>> port).  I know the server is working because I can see the web interface
>>> if I view http://optimus:8080 either locally on the server or remotely
>>> from my laptop.
>>>
>> [snip]
>>
>>> DEBUG [main] (SessionConnection.java:273) - Finding session factory for rmi://optimus/server1
>>>
>> [snip]
>>
>>> Caused by: java.net.ConnectException: Connection refused: connect
>>>
>> [snip]
>>
>>>     at java.net.Socket.<init>(Socket.java:179)
>>>     at sun.rmi.transport.proxy.RMIDirectSocketFactory.createSocket(RMIDirectSocketFactory.java:22)
>>>
>> [snip]
>>
>> Note that RMI uses several ports, not just 1099 one. Specifically,
>> it'll connect to registry on 1099 first to get the client stubs, but
>> then it'll use random ports to actually do the RMI calls. Now, you say
>> you've disabled firewalls, does that mean you have _all_ ports (all
>> the way up to 65000) open?
>>
>> A couple suggestions:
>>
>>  1. Try starting mulgara with '--rmiobjectport 1099'  (i.e. can be the
>>     same port as used for the registry) - this will cause the actual
>>     RMI calls to use port 1099 too.
>>
>>  2. Fire up wireshark (or whatever your favorite snooper is) and see
>>     what it's trying to connect to that is failing.
>>
>>
>>   Cheers,
>>
>>   Ronald
>>
>>
> _______________________________________________
> Mulgara-general mailing list
> Mulgara-general at mulgara.org
> http://mulgara.org/mailman/listinfo/mulgara-general
>