condor-remote-configuration-server package. Only one machine in a cluster should act as the server, and should be the same machine that the MRG Management Console is installed on. Install the package on the server using yum:
# yum install condor-remote-configuration-server
CNAME record will need to be added to the DNS configuration with the name puppet. It needs to point to the machine that has the condor-remote-configuration-server installed.
puppet.conf.master file to the /etc/puppet directory using the following command:
$ cp -f /etc/opt/grid/examples/puppet.conf.master /etc/puppet/puppet.conf
site.pp file:
$ echo 'import "condor"' >> /etc/puppet/manifests/site.pp
# service puppetmaster start Starting puppetmaster service: [ OK ]
condor-remote-configuration package. Install the package using yum:
# yum install condor-remote-configuration
puppet.conf.client file to the /etc/puppet directory using the following command:
$ cp -f /etc/opt/grid/examples/puppet.conf.client /etc/puppet/puppet.confThen copy the
namespaceauth.conf file to /etc/puppet:
$ cp -f /etc/opt/grid/examples/namespaceauth.conf /etc/puppet
namespaceauth.conf configuration file in your preferred text editor and locate the line that states allow <puppetmaster.fqdn> entry. Replace the <puppetmaster.fqdn> text with the fully qualified domain name of the server machine. This is the machine that has the condor-remote-configuration-server package installed:
allow server.example.com
# service puppet start Starting puppet service: [ OK ]
$ condor_configure_node -n [name of node] [action] [features]Use the
--help option to see a full list of possible commands:
$ condor_configure_node --help
--add|-a: Used to add features to the node being configured. If any additional information is required, the tool will prompt for it.
--delete|-d: Used to remove features on the node.
--list|-l: Used to list configurations. If provided without a specified node, this will print the list of nodes being managed. If a node is specified, it will print the list of features for that node. If provided with one or more features in addition to a node, then it will print the specific configurations for those features.
Schedulers: If the node being configured is not a scheduler, then the configuration tool will prompt for the list of schedulers the node should be allowed to submit to. First it will prompt for the default scheduler, then for a comma separated list of additional schedulers. This entry should contain the fully qualified domain names of the schedulers in the pool. If the pool has a high availability scheduler, use ha-schedd@.
Collector Name: This is a human readable text field that will identify the pool. This value will be set as the COLLECTOR_NAME for the node.
QMF Broker Information: The QMF Broker is the AMQP broker that is used to communicate with the MRG Management Console. The configuration tool will prompt for the IP or hostname where the broker is running, as well as the port the broker is listening on. If no port is provided, the default port will be used.
~/condor_config.local local configuration file. It is possible to provide a custom configuration for a node by creating a file named ~/condor_config.overrides and adding configuration entries to that file.
~/condor_config.overrides file. Settings in this file will override any settings provided by the remote configuration feature. This can result in lost or incorrect functionality of the features controlled by the remote configuration feature.
condor_ca.domain.com to be a High Available Central Manager:
$ condor_configure_node -n condor_ca.domain.com -a -f ha_central_manager
twofer.domain.com to be both a scheduler and an execute node:
$ condor_configure_node -n twofer.domain.com -a -f scheduler,startd
twofer.domain.com:
$ condor_configure_node -n twofer.domain.com -d -f startd
$ condor_configure_node -l
twofer.domain.com:
$ condor_configure_node -l -n twofer.domain.com
twofer.domain.com:
$ condor_configure_node -l -n twofer.domain.com -f ha_central_manager