[AG-TECH] AG2.0 and linux

[Thomas Uram]

Dave:

CVS is our live code base, so it's always changing.  It has changed 
since beta3 was released.  Using CVS code has implicit caveats because 
it is development code, which means it is unsupported and difficult to 
report bugs against.

Having said that...

If you want to connect to the vv2:9000 server, you should use the venue 
client included in the beta3 release.

If you want to run a venue server, you can execute VenueServer.py.  By 
default, it will run on hostname:8000.  The venue server will create a 
single venue at startup.  To access this server from outside your 
firewall, you would need to open ports 8000, 8002, 8004, and 8006.  We 
are working on user guide materials now, so you'll have a better 
reference in the future.

The client/server mismatch is a problem simply because the venuestate 
data structure changed between beta3 and now.  vv2:9000 is running 
beta3, and you are running a post-beta3 client, so they each assume a 
different venue state data structure.

I hope this helps.

Tom



Dave Semeraro wrote:

>
> Hello Tom,
>
> So I was under the impression that the CVS code was the latest release 
> ( ie beta3). Is that not
> the code source that your server is running from? Randy Heiland 
> downloaded a "beta3" rpm and
> keeps getting invalid URL errors when he tries to connect to the 
> venueserver (vv2...).
>
> I am a little confused, is the client code in CVS pre beta3 or post 
> beta3? Should the venue client
> expect the services and clients in the venue state or not? Is the 
> VenueClient code or your
> server doing it right? Are there instructions somewhere on how to run 
> a server?
>
> Dave
> At 05:34 PM 4/23/2003 -0500, you wrote:
>
>> Hi Dave:
>>
>> Dave Semeraro wrote:
>>
>>>
>>> Hello AG-TECHers,
>>>
>>> Well after a couple of days of frustration I finally got the 
>>> VenueClient to connect to the ANL
>>> venue. For reference my software state is the CVS update as of 
>>> 5:35pm on April 22. I am
>>> sure things have changed by now.
>>>
>>> My problem was that under linux I could not get the RunMe.py script 
>>> to connect. At first I
>>> was getting invalid url errors but that went away after a CVS update 
>>> or two. Finally I was getting
>>> python errors saying that venueState had no attribute called 
>>> clients. The VenueClient would then
>>> hang. I commented out the "for client in venueState.clients" loop 
>>> and had to comment out the
>>> "for service in venueState.services" loop in the VenueClient.py file 
>>> in the
>>> AccessGrid/AccessGrid directory. ( all this will make sense to those 
>>> of you who have messed
>>> with the code). That allowed the VenueClient to function.. sort of. 
>>> I had nothing under services or
>>> users in the gui. Probably because I commented that part out of the 
>>> code.
>>>
>>> The venueState object
>>> gets initialized in a call to 
>>> "Client.Handle(URL).get_proxy().Enter(self.profile)". For some reason
>>> the clients data and services data are not set as a result of this 
>>> call. This call is found in the
>>> VenueClient.py file in the EnterVenue method. My head is starting to 
>>> swim with all the code
>>> tracing. Can anyone give me a clue as to why I have no clients or 
>>> services returned? I probably
>>> have a configuration file missing or something.
>>
>>
>> This problem is a version mismatch between the venue server and venue 
>> client.  The venue client code you have is expecting clients and 
>> services in the venue state, but the server to which you are 
>> connecting (presumably vv2:9000)  is running beta3 code which does 
>> not include these in the venue state.  If you run a client and server 
>> from the same version of the code (both from beta3, say), they would 
>> be using the same definition of venue state, and this problem 
>> wouldn't occur.  You could do the same using cvs code, but then there 
>> are fewer guarantees.
>>
>>> I was able to upload an image file to the venue but not download it. 
>>> Anyone have any idea of what I may be facing on that problem?
>>
>>
>> If you can provide details on how to reproduce the problem, or text 
>> from exceptions that you saw, or .log files that captured details of 
>> the problem, we can debug it.  Ideally, of course, this info would be 
>> included with a bug report in bugzilla.
>>
>> Thanks,
>>
>> Tom
>>
>>
>
> Dave Semeraro Ph.D.
> Visualization and Virtual Environments Group
> NCSA University of Illinois
> 605 E. Springfield Ave.
> Champaign, IL 61820
> Semeraro at ncsa.uiuc.edu
> (217) 244-1852
>
>