[Condor-users] Some networking oddness.

[Andrew Walker <amw75@xxxxxxxxx>]

Hi all,

I recently created a new pool using (linux X86_64) machines with two network interfaces (only one being active) and spotted a couple of oddities in the way condor (6.8.4) behaves in what are probably rather unusual circumstances. This information may be useful to somebody else and prevent head scratching in the future so I have included it below.

The first issue seems to arise as (for reasons that are far too boring to go into) the active interface on all the machines is eth1 (lo is also active but eth0 is not). In this case starting condor results in all the daemons binding to the loopback address rather than eth1. Adding NETWORK_INTERFACE = xx.xx.xx.xx to the config file acts as a workaround and the daemons bind to the correct interface. 

The second issue involves changing HIGHPORT and LOWPORT in the configuration file and using condor_reconfig.  Daemons do not appear to change ports if they are bound to ports outside the new range. Stopping and restarting all the daemons results in them being bound to a correct port. I note that for some configuration variables section 3.3 of the manual states "This configuration variable cannot be changed by using condor_ reconfig or by sending a SIGHUP", but this is not true of any of the network-related entries. Perhaps such a warning should be added. 

Cheers,

Andrew

--

Dr Andrew Walker

Department of Earth Sciences

University of Cambridge

Downing Street

Cambridge 

CB2 3EQ

UK