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

[HTCondor-users] Networking of the Worker-nodes in Flocks


I am a bit confused about how condor flocking works. 

I have two clusters C1 and C2. Resources in C2 are used with the Flocking. C1 and C2 are in different network. C1 main node schedule incoming jobs first to the C1 available resources and once all get filled, it start utilizing the C2 resources. 
For my initial experimental setup C2 is based on one machine and that have a live IP. Now I want add more nodes to C2 flock. But it seems like each node requires Live IP. With private IP, nodes in C2 changed their 
status to "Matched" but jobs never started on them whereas the node with Live IP runs the job without any problem.

In my initial understanding, only the main-node in C2 should have live-ip and all the worker nodes should have internet access but live-ip is not required. But that seem not the case... 

Now my question is do I really need a live IP for every node in C2? If yes than what the point of making the flock? Why not I directly attach each node of C2 cluster to the main node of C1? 

I think I am missing some concept here. Can someone please explain this to me or refer some guide?