LIGO Data Grid

CondorView is a Condor contribution module that can be used to provide statistical information on resources and usage of a Condor pool. These information (e.g. user-based resource usage statistics) cannot be obtained from low-level cluster monitoring tools (e.g. Ganglia).

It will be very helpful to know briefly how CondorView works before going to configuration steps. A Condor pool central manager runs the main condor_collector and writes Condor usage information into log files. A CondorView server is a machine within the Condor pool that runs another condor_collector daemon. The CondorView condor_collector recieves information dispatched periodically from the main condor_collector and writes them into history files. The CondorView client is any machine that runs a web server and can reach the CondorView server. The CondorView client runs cron jobs that contact periodically the CondorView server to get up-to-date information and create corresponding html pages. In summary, CondorView works in this way:

log files (central manager) -> history files (CondorView server) -> html files (CondorView client)

Go to here to configure and deploy a Condor pool. With an existing Condor, follow steps below to configure and deploy CondorView:

1. Configure a machine in the Condor pool to be a CondorView server by adding two settings to the local configuration file of the chosen machine:

   POOL_HISTORY_DIR = /export/condor-history
   

   KEEP_POOL_HISTORY = True
   

2. Configure Condor to automatically start the CondorView server by defining:

   VIEW_SERVER = $(SBIN)/condor_collector
   

   DAEMON_LIST = MASTER, STARTD, SCHEDD, VIEW_SERVER
   

3. Configure a pool to report to the CondorView server by defining the following setting in the global configuration file:

   CONDOR_VIEW_HOST = full.hostname
   

   where full.hostname is the full hostname of the CondorView server. If you do not have a shared global configuration file for Condor, you should put the same value in the configuration files on both the main condor_collector and the CondorView server host.
4. Restart the condor_master at your CondorView server host.
5. Send a condor_reconfig to your main condor_collector for the change to take effect.
6. It will take a while for the CondorView server to produce history files. Make sure those files generated in POOL_HISTORY_DIR are not all empty. Otherwise, check for error messages in the Condor collector log files of both main and CondorView server hosts.
7. Choose a machine with a web server running as the CondorView client.
8. Go to here to download the CondorView client module.
9. Fill in the customer information and click on "I Agree".
10. The Condor contributed module download page includes a huge list of modules. Scroll down to the end of the page and download the tar.Z file of the Version 6.1.8 (view_client).
11. Unpack the CondorView client package by running:

    gunzip view_client-6.1.8-Any-Java.tar.Z
    

    tar xvf view_client-6.1.8-Any-Java.tar
    

    cd view_client-6.1.8
    

12. Follow the instructions in the file INSTALL to configure and deploy the CondorView client. This is a very well-written guide.
13. Have your web browser Java enabled to view CondorView pages.

This completes a basic deployment and configuration of CondorView. Please go to the Condor manual for details. Any problems, please contact Junwei Cao.

Supported by the National Science Foundation. Any opinions, findings and conclusions or recomendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF)

---

$Id: condorview.html,v 1.16 2007/11/06 03:41:41 patrick Exp $