Yes, that is expected. The negotiator and collector daemons should only be running on one machine (which we call the Central Manager). What value is the parameter ALLOW_NEGOTIATOR set to in the Condor configuration file on your compute nodes?
yes, they are on the same machine. But I see this log on my compute nodes, in StartLog file, and collector is not running on them. Is this ok?
2014-3-6 AM6:37ä "Jaime Frey" <jfrey@xxxxxxxxxxx
On Feb 24, 2014, at 1:53 AM, éææ <kyleqian@xxxxxxxxx
I find this log entry in StartLog file:
02/24/14 10:29:46 slot4: State change: claiming protocol successful
02/24/14 10:29:46 slot4: Changing state: Unclaimed -> Claimed
02/24/14 10:29:46 PERMISSION DENIED to unauthenticated@unmapped from host 10.1.1.1 for command 440 (MATCH_INFO), access level NEGOTIATOR: reason: NEGOTIATOR authorization policy contains no matching ALLOW entry for this request; identifiers used for this host: 10.1.1.1,amax.local, hostname size = 1, original ip address = 10.1.1.1
the last log entry will repeat many times for every slot. It seems that some settings for negotiator are not correct. What should I do to correct this?
This is controlled by the ALLOW_NEGOTIATOR configuration parameter. The default setting allows connections from the machine running the condor_collector daemon. Are your condor_negotiator and condor_collector running on the same machine (the usual case)? Is that machine 10.1.1.1?
UW-Madison HTCondor Project