catalog_server(1) | Cooperative Computing Tools | catalog_server(1) |
catalog_server - start a catalog server
catalog_server [options]
A catalog server provides naming and discovery for multiple components of the Cooperative Computing Tools, particularly the Chirp distributed filesystem and the Work Queue distributed programming framework. Services that wish to be known on the network periodically publish their information to the catalog server via a short UDP packet. Clients wishing to discover services by name may query the catalog by issuing an HTTP request to the catalog server and will receive back a listing of all known services.
To view the complete contents of the catalog, users can direct their browser to http://catalog.cse.nd.edu:9097. Command line tools work_queue_status and chirp_status present the same data in a form most useful for Work Queue and Chirp, respectively. Large sites are encouraged to run their own catalog server and set the CATALOG_HOST and CATALOG_PORT environment variables to direct clients to their server.
The catalog server is a discovery service, not an authentication service, so services are free to advertise whatever names and properties they please. However, the catalog does update each incoming record with the actual IP address and port from which it came, thus preventing a malicious service from overwriting another service's record.
On success, returns zero. On failure, returns non-zero.
The Cooperative Computing Tools are Copyright (C) 2003-2004 Douglas Thain and Copyright (C) 2005-2015 The University of Notre Dame. This software is distributed under the GNU General Public License. See the file COPYING for details.
CCTools 7.0.9 FINAL |