[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] How to Test if flocking is working?



Thanks! after using

condor_status -pool flocked_pool -name central_node_in_flocked_pool

we were able to see eachother's machines..now trying to get condor_q to work..thanx a bunch for the tip.

JW



Kewley, J (John) wrote:

Are there some known difficulties with firewalls even if the ports are opened and forwarded to the proper machines within the network??



As long as ALL submit machines are able to talk to ALL execute machines and the central node, firewalls shouldn't be a problem. Of course this means all of the following: * No private networks (ie if there were no firewalls, machines could still talk) * All machine FWs have 9614, 9618 and user-defined upper port range opened for TCP AND UDP, and the upper port range is declared in the condor config files. * Any department or site firewalls which are 2in the way" are also opened in the same way



So far the other pool doesn't show up in my condor_status output and none of my jobs indicate that they ar going go machines that are not on my local pool.



try the following: condor_status -pool flocked_pool (from both sides of the flock) condor_q -global condor_q -pool flocked_pool -name central_node_in_flocked_pool condor_q -pool flocked_pool -name other_node_in_flocked_pool

You can also try submitting directly to the central node in the flocked pool,
by specifying Machine or Name in the REQUIREMENTS statement.

There are also programs you can run to check which ports are open on machines,
but use of these are frowned upon in many institutions!

JK