[HTCondor-users] Negotiator Problems

["Thomas, Dallas" <Dallas.Thomas@xxxxxxxxx>]

Greetings All,

 

I have been testing Condor both without a NFS mount and now with a NFS mount.  After creating the NFS mount – I left the condor_config file for the central manager and the single (startd) node in /etc/condor.  And have them point to the nfs mount for the condor_config.local which I have called condor_config.cluster that then points to each condor_config.$(HOSTNAME).

 

Things appeared to be working fine until I tried running a job which proceeded to remain idle.  After some checking through the logs I discovered inside the CollectorLog the following error:

 

06/11/14 08:32:36 PERMISSION DENIED to unauthenticated@unmapped from host 127.0.0.1 for command 10 (QUERY_STARTD_PVT_ADS), access level NEGOTIATOR: reason: cached result for NEGOTIATOR; see first case for the full reason

 

Now my collector, negotiator and schedd are all on the central manager.  I figured this might have something to do with the ALLOW_NEGOTIATOR in the condor_config on the CM.  I have set it to $(CONDOR_HOST) and I have left it blank.  Neither has had any effect on the error in the CollectorLog.

 

The condor_config for each CM and Node have virtually remained untouched from the install via dpkg install.  The condor_config.cluster/headnode/worker are below.

 

condor_config.cluster:

 

LOCAL_CONFIG_FILE = /scratch/nfs/condor/condor-etc/condor_config.$(HOSTNAME)

 

UID_DOMAIN = local

 

COLLECTOR_NAME = "Debian Condor Pool at $(UID_DOMAIN)"

 

FILESYSTEM_DOMAIN = $(UID_DOMAIN)

CONDOR_ADMIN = root@$(FULL_HOSTNAME)

 

CONDOR_HOST = ablethr21.local

 

IN_HIGHPORT = 9999

IN_LOWPORT = 9000

 

#SEC_DAEMON_AUTHENTICATION = required

#SEC_DAEMON_AUTHENTICATION_METHODS = password

#SEC_CLIENT_AUTHENTICATION_METHODS = password,fs,gsi

#SEC_PASSWORD_FILE = /var/lib/condor/condor_credential

#ALLOW_DAEMON = condor_pool@*

#ALLOW_NEGOTIATOR = condor_pool@$(UID_DOMAIN)/$(CONDOR_HOST)

#SEC_DEFAULT_AUTHENTICATION_METHODS = FS

 

NEGOTIATOR_INTERVAL = 20

 

TRUST_UID_DOMAIN = TRUE

 

START = TRUE

SUSPEND = FALSE

PREEMPT = FALSE

KILL = FALSE

 

condor_config.headnode:

 

DAEMON_LIST = MASTER, COLLECTOR, SCHEDD, NEGOTIATOR

ALLOW_WRITE = $(ALLOW_WRITE), 10.115.0.56,ablethr26.local

 

condor_config.worker:

 

DAEMON_LIST = MASTER, STARTD

ALLOW_WRITE = $(ALLOW_WRITE), $(CONDOR_HOST)

 

Any help would be very much appreciated.

 

Sincerely

Dallas Thomas