Well, the ports that get opened up are by a new container started for
the purpose of consuming notifications. The GlobusRun client might
not be smart enough to pass those properties along to the container
that gets started, but I believe it's not an officially supported
client. The Java API should be able to handle the port ranges
correctly, so long as you make sure that the notification consumer is
getting started with the right settings.
If you're so inclined, you could probably get this to work by
creating a ~/.globus/cog.properties file with the right port range in
it. That would also get read in by the notification consumer
container. Otherwise, if you want to fix it at the installation
level, you could add some port range options to the $GLOBUS_LOCATION/
etc/globus_wsrf_core/client-server-config.wsdd file, which controls
the options used by all notification-consuming containers.
Charles
On Nov 25, 2007, at 9:13 PM, Ryan Zhang wrote:
Hi,
I am using the java API to submit gramJob from site A(client)
to site B(server) and I have open the Port from 50000 to 51000 on
both site, however I can't get job status notification. The
globusrun-ws works fine, here are the output
[EMAIL PROTECTED]:~> globusrun-ws -submit -F https://
gridgate.rtc.rice.edu -c /bin/sleep 60
Submitting job...Done.
Job ID: uuid:9d246290-96cc-11dc-8639-000eab01fd60
Termination time: 11/20/2007 18:24 GMT
Current job state: Active
Current job state: CleanUp
Current job state: Done
Destroying job...Done.
(client is the gridgate.ada.rice.edu and the server is the
gridgate.rtc.rice.edu)
on rtc(server) I tried
-bash-3.00$ netstat | grep ada
tcp 0 0 gridgate.rtc.rice.edu:8443 gridgate.ada.rice.edu:
50001 TIME_WAIT
tcp 0 0 gridgate.rtc.rice.edu:56467 gridgate.ada.rice.edu:
50000 TIME_WAIT
which is exactly what it should do!
However, if I tried the Java globusRun client here is the command
that should do the same thing but it doesn't
[EMAIL PROTECTED]:~> java -DGLOBUS_TCP_PORT_RANGE=
$GLOBUS_TCP_PORT_RANGE -DGLOBUS_TCP_SOURCE_RANGE=
$GLOBUS_TCP_SOURCE_RANGE -DGLOBUS_LOCATION=$GLOBUS_LOCATION
org.globus.exec.client.GlobusRun -factory https://
gridgate.rtc.rice.edu:8443/wsrf/services/ManagedJobFactoryService "/
bin/sleep 60"
Submission ID: uuid:fbd5ba70-96d9-11dc-ae18-d1f204ba9e67
WAITING FOR JOB TO FINISH
2007-11-19 14:00:51,549 WARN client.GlobusRun
[main,waitForJobCompletion:1199] Did not receive any new notification
of job state change after a delay of 60000 ms.
Pulling job state.
========== State Notification ==========
Job State: Active
========================================
2007-11-19 14:03:51,976 WARN client.GlobusRun
[main,waitForJobCompletion:1199] Did not receive any new notification
of job state change after a delay of 60000 ms.
Pulling job state.
========== State Notification ==========
Job State: Done
========================================
Exit Code: 0
DESTROYING JOB RESOURCE
JOB RESOURCE DESTROYED
and on the server side I tried
-bash-3.00$ netstat | grep ada
tcp 0 0 gridgate.rtc.rice.edu:8443
gridgate.ada.rice.edu:40394 TIME_WAIT
The port is not in range.
I wonder if there is anyway to correct this? Thanks.
-Ryan
You keep typing, we keep giving. Download Messenger and join the
i’m Initiative now. Join in!
