[HTCondor-users] newbie ask a simple question about condor installation. Thank you very much!

[btdan <btdan@xxxxxxx>]

Hi,All.

 I have begin to learn condor. And I get a problem with HTCondor. Can somebody be kind enough to help me solve this problem ?Thank you very much.

And I install condor 7.9.4 on my Centos System as follows:

 after I get htcondor-stable-rhel6.repo  and htcondor-development-rhel6.repo ,I install condor simply with the following command:

 

# yum install condor.x86_64

and then I can start condor service:

#service condor start

and check whether condor is running

#ps aux |grep condor

And It is running normally. I can submit a job and get the result correctly.

 

And then I install condor on another Server with Centos System. And it can run normally.

But The question is : How to make both Server connected as a simple cluster?

I run the command

#condor_configure --owner=myowner --type=submit,execute --central-manager=158.26.18.202

and the result is :

------------------ The result begin -------------------

Usage: host [-aCdlriTwv] [-c class] [-N ndots] [-t type] [-W time]

            [-R number] [-m flag] hostname [server]

       -a is equivalent to -v -t ANY

       -c specifies query class for non-IN data

       -C compares SOA records on authoritative nameservers

       -d is equivalent to -v

       -l lists all hosts in a domain, using AXFR

       -i IP6.INT reverse lookups

       -N changes the number of dots allowed before root lookup is done

       -r disables recursive processing

       -R specifies number of retries for UDP packets

       -s a SERVFAIL response should stop query

       -t specifies the query type

       -T enables TCP/IP mode

       -v enables verbose output

       -w specifies to wait forever for a reply

       -W specifies how long to wait for a reply

       -4 use IPv4 query transport only

       -6 use IPv6 query transport only

       -m set memory debugging flag (trace|record|usage)

WARNING: Unable to determine full hostname for host 'node2'. Condor may not work properly

Please set the environment variable $HOSTNAME to the full name of this machine (e.g., mymachine.mydomain.com)

Directory /home/lijian/condor/bin does not exist

 

Configured condor using these configuration files:

  global: /etc/condor/condor_config

  local:  /etc/condor/condor_config.local

 

In order for Condor to work properly you must set your CONDOR_CONFIG

environment variable to point to your Condor configuration file:

/etc/condor/condor_config before running Condor commands/daemons.

Created scripts which can be sourced by users to setup their

Condor environment variables.  These are:

   sh: /usr/condor.sh

  csh: /usr/condor.csh

 

------------------ The result end -------------------

 

and then :

#sh /usr/condor.sh

#service condor restart

 

However ,nothing is changed. With the command condor_q,I can see :

 

 

-- Submitter: node2 : <158.26.18.203:43554> : node2

 ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD

   2.0   myowner          3/22 15:21   0+00:00:06 I  0   0.0  java myComput

 

1 jobs; 0 completed, 0 removed, 1 idle, 0 running, 0 held, 0 suspended

 

the 158.26.18.203 is its own IP address.

Can somebody be kind enough to help me solve this problem ?

Thank you very much.

Best wishes.

 

---