The Distributed Parallel Storage Server (DPSS) User Guide

Lawrence Berkeley National Laboratory, Data Intensive Distributed Computing Project

This page contains instructions for using the DPSS and it's various support programs.

Starting the DPSS:

To use the DPSS, your application must connect to the DPSS and the DSM (Data Set Manager). The DSM is needed to figure out which "set id" you want to get, and contains all data set specific information, such as data set name, size, type, etc. To start the DPSS and DSM, just type: 

 dpss_start hostname
dpss_start will start a master and server on a given host. To start a server only, type: 
 dpss_serv_restart hostname
Make sure the correct values for DPSSHOST, DPSSHOME, and USER are in the control script 'dpss_control' before doing running these scripts.

There are two shell environment variables that should be set when running any DPSS program from the command line. These are: DPSSHOST and DPSSHOME. For example: 

	setenv DPSSHOME /usr/local/DPSS
	setenv DPSSHOST dpss.lbl.gov

To verify the DPSS is working, try running 'dpss_status'. If you get any error messsage, check the file $DPSSHOME/logs/*.log for error messages. If dpss_status works, then you know that both the DPSS and DSM are up and accepting connections.

Killing the DPSS:

If you need to kill off the DPSS master and servers, use the 'dpss_stop' script. The DPSS uses semaphores and shared memory, which might not get freed up properly if you kill the DPSS processes off by hand. The dpss_stop script ensures all semaphores and shared memory are properly cleaned up. The stop script are used as follows: 

 dpss_stop hostname

Loading Data Sets:

The simplest way to load a file to the DPSS is using the program dpss_put . For Image tiles, the program loaddata is used instead. For more information on how to load TerraVision data to the DPSS, go to the Data Loading Instructions Page .

Useful DPSS support programs:

dpss_ls: List all sets defined on the DPSS. Use "dpss_ls -v" to get detailed information on the data set, and "dpss_ls -D" to find out which servers a set is loaded on.

dpss_status: Check the status of the DPSS master.

Note the these programs all use the environment variable DPSSHOST, or command line options, to set DPSS hostnames.

COPYRIGHT NOTICE: Distributed Parallel Storage System and DPSS/ISS are Copyright © 1994 by Lawrence Berkeley Laboratory

This page is the resonsiblity of Brian L. Tierney. (BLTierney@lbl.gov). Support Credits are here. This document and its uses are subject to LBL's disclaimers and legal notices.

Credits: The research and development of the Distributed Systems Department i s funded by the U.S. Dept. of Energy, Office of Science, Office of Advanced Scie ntific Computing Research, Mathematical, Information, and Computational Sciences Division

If there is a problem with this page please, e-mail webmaster@www-itg.lbl.gov.

This page last modified: [an error occurred while processing this directive]