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

[Condor-users] Condor manager at a frontend machine with two NICs



Hi

I'm setting up a Windows cluster and installing the Condor. It has a frontend node having two NICs - one for private and one for public. For the private network, it is given 192.168.0.1 and all other compute nodes with 192.168.0.* are connected via a switch. The public network interface has setup with a static IP, a gateway and a DNS Server to allow connections to and from the outside network (Intenet) -- pretty much a basic setup for a Beowulf cluster.

The problem is that if the public network is enabled, condor_status gives nothing and condor_q give an error saying that <192.168.0.1:xxx> is not found or something like that. If I do "condor_status" on compute nodes, they correctly list the Condor pool with all nodes slots on it. If the NIC for public network is disabled, then the Condor works from the frontend node; I can submit a job with condor_submit and see the resource list with condor_status with no problem.

Is there any special Condor configuration needed for the frontend manager node with two NICs? I read some email threads regarding two NICs, mentioning about BIND_ALL or something like that to be specified in the config file. But I'm not sure what it is for and if that solved the problem. Would anyone help me on this issue? Any help would be appreciated. Thank you.

Best,

Jeho