Hello,
I apologize if this has been addressed else where. I searched
the archive and didn't find the answer. My group is in the process
of upgrading to 3.1 from 2.4. For various political reasons I've
been asked to keep a 2.4 server running for the near future. I've
had a 2.3/2.4 Venue/Bridge server operational under RHEL3 for over 3
years now and am quite comfortable with setting up and maintaining
2.X AG.
We've acquired new hardware for the 3.1 Venue/Bridge setup
running RHEL5 and I wanted to run both the 2.X and 3.X servers on
the same box to make my administrative life easier. Using the docs
from:
http://www.vislab.uq.edu.au/research/accessgrid/software/rhel/
I installed all the AG3.1 RPM's (and pre-requirements) as well as
the 2.4 compatibility set. I have the new 2.4 Venue Server up and
operational and working fine from 2.X clients, but I can not get the
bridge server to connect to the new Venue Server correctly.
Logs on the new bridge server when it connects to the new venue
server show:
05/21/08 09:43:33 46912496258912 Toolkit Toolkit.py:709 INFO
Service init: have profile None
05/21/08 09:43:33 46912496258912 Toolkit Config.py:214 DEBUG
System hostname of bridge3.renci.org is valid
05/21/08 09:43:33 46912496258912 CertificateManager
CertificateManager.py:269 DEBUG Opened repository
/opt/ag/.AccessGrid/Config/certRepo
05/21/08 09:43:33 46912496258912 Toolkit Toolkit.py:738 INFO
Initialized cert mgmt.
05/21/08 09:43:33 46912496258912 Toolkit Toolkit.py:753 INFO
Loaded profile and configured with it.
05/21/08 09:43:33 46912496258912 CertificateManager
CertificateManager.py:759 DEBUG Configuring standard environment
05/21/08 09:43:33 46912496258912 CertificateManager
CertificateManager.py:827 DEBUG Using default identity /O=Access
Grid/OU=agdev-ca.mcs.anl.gov/CN=VenueServer/venues.renci.org
05/21/08 09:43:33 46912496258912 CertificateManager
CertificateManager.py:1083 DEBUG Initializing environment with
unencrypted cert /O=Access
Grid/OU=agdev-ca.mcs.anl.gov/CN=VenueServer/venues.renci.org
05/21/08 09:43:33 46912496258912 CertificateManager
CertificateManager.py:1572 DEBUG done, success=1
05/21/08 09:43:33 46912496258912 Toolkit Toolkit.py:764 INFO
Initialized Globus.
05/21/08 09:43:33 46912496258912 Toolkit Toolkit.py:782 INFO
Service Initialization Complete.
05/21/08 09:43:33 46912496258912 Toolkit BridgeServer24:157 INFO
BridgeFactory.SetPortMin 30000
05/21/08 09:43:33 46912496258912 Toolkit BridgeServer24:161 INFO
BridgeFactory.SetPortMax 30999
05/21/08 09:43:33 46912496258912 Toolkit BridgeServer24:259 INFO
AddVenueServer: url = https://venues.renci.org:9000/VenueServer
05/21/08 09:43:33 46912496258912 Toolkit BridgeServer24:274 INFO
AddVenue: url =
https://venues3.renci.org:9000/Venues/00000102403861290098001300a80081d1e
05/21/08 09:43:34 46912496258912 EventClient EventClient.py:158
DEBUG Have callback handle _303002b4aa2a0000_p_callbackStruct
05/21/08 09:43:34 1115699520 Toolkit BridgeServer24:612 INFO
Method Venue.RunQueueThread called
05/21/08 09:43:34 46912496258912 Toolkit BridgeServer24:274 INFO
AddVenue: url =
https://venues3.renci.org:9000/Venues/00000102403832e40098001300a8008150a
05/21/08 09:43:35 46912496258912 EventClient EventClient.py:158
DEBUG Have callback handle _30c5be1900000000_p_callbackStruct
05/21/08 09:43:35 1147169088 Toolkit BridgeServer24:612 INFO
Method Venue.RunQueueThread called
05/21/08 09:43:35 46912496258912 Toolkit BridgeServer24:274 INFO
AddVenue: url =
https://venues3.renci.org:9000/Venues/000001024035bd6c0098001300a80081bfc
05/21/08 09:43:36 46912496258912 EventClient EventClient.py:158
DEBUG Have callback handle _d0c2c21900000000_p_callbackStruct
05/21/08 09:43:36 1178638656 Toolkit BridgeServer24:612 INFO
Method Venue.RunQueueThread called
05/21/08 09:43:36 46912496258912 Toolkit BridgeServer24:274 INFO
AddVenue: url =
https://venues3.renci.org:9000/Venues/000001024037fb040098001300a80081610
05/21/08 09:43:37 46912496258912 EventClient EventClient.py:158
DEBUG Have callback handle _50c101b4aa2a0000_p_callbackStruct
05/21/08 09:43:37 1210108224 Toolkit BridgeServer24:612 INFO
Method Venue.RunQueueThread called
05/21/08 09:43:37 46912496258912 Toolkit BridgeServer24:274 INFO
AddVenue: url =
https://venues3.renci.org:9000/Venues/000001024a6692130098001300a80081582
05/21/08 09:43:38 46912496258912 EventClient EventClient.py:158
DEBUG Have callback handle _20b800b4aa2a0000_p_callbackStruct
05/21/08 09:43:38 1241577792 Toolkit BridgeServer24:612 INFO
Method Venue.RunQueueThread called
The configuration file for the Bridge Server is:
[BridgeServer]
name = RENCI3
location = EUROPA
qbexec = /usr/bin/QuickBridge
portMin = 30000
portMax = 30999
[https://venues3.renci.org:9000/VenueServer]
type = VenueServer
Everything looks correct, the bridge connects to the server,
gets all the Venues, but QuickBridge never starts up. If I point
this same bridge server at my old 2.4 server (under RHEL3), it
establishes a connection, and will serve Unicast just fine, so I
don't think its the bridge server. I already dropped the firewall
on both the Venue and Bridge server just to be sure it wasn't a port
issue.
I have 2 things I think it could be:
1) conflict with certificate name. I am using the certificate from
my old server (venues.renci.org) on this test server
(venues3.renci.org). Thinking it might be a naming conflict, I
renamed the new server "venues.renci.org" and that didn't solve the
problem, and I see no complaints in the logs about certificate issues.
2) The newer wxPython (2.8.7.1) and associated pre-requirements for
AG 3.1 (downloaded from EPEL) are too new to correctly work with the
AG 2.4 setup. I ran into a couple of issues trying to start the 2.4
Client, problems with wxPython compatability, but nothing on the
server side, and everything else on the 2.4 Venue Server is working
fine as far as Multicast (I can connect from a 2.4 clients and have
sessions just fine), its just the Unicast bridge won't connect and
start the QuickBridge for any of the Venues.
Before I spend the time to rebuild the software stack myself
using older code on RHEL5 (wxPython 2.5.x, etc), I thought I would
ask if anyone has experienced this before, and knows of a quick
solution or can offer a direction for me to look in.
Thanks,
-Brad Viviano