STAR   Computing  GRID  Monitoring  MonALISA

The MonALISA Pseudo-Client Deployment

Introduction

The MonALISA pseudo-client retrieves monitoring data from MonALISA services that are members of one or more monitoring groups. The pseudo-client connects to the monitoring services via proxy servers. There are currently four proxy servers available. The monitoring data collected by the pseudo-client are cached locally and are made available to other applications that may have use for such data. The pseudo-client code runs on Web Repositories, caches the retrieved data which are then stored in a database. The pseudo-client was made available to us by the MonALISA developers and we have deployed it to cache monitoring data from sites of interest locally.

Getting the Pseudo-Client code

The pseudo-client code consists of the server code and the client code. The server starts a thread that collects the monitoring data from the various sites and listens for incoming connections to port 9331. When a connection is established to this port the server spawns another thread that serves the monitoring data. The client code simply establishes a connection to the server and gets the latest cached monitoring data. The hostname of the server and the port number are hardwired in the client code. Firewall requirements: outgoing traffic in the range 6000-6003 tcp to the four proxy servers must be open in order the server to collect the monitoring data. The pseudo-client code is available from the BNL CVS Repository. To get the code, first, You need to set your CVSROOT environment variable, and then login: 
              $export CVSROOT=:pserver:cvsuser1@cvs.bnl.gov/data01/CVS
              $cvs login
                Logging in to :pserver:cvsuser1@cvs.bnl.gov:2401/data01/CVS
                CVS password:
The password is the same as the CVS username: cvsuser1. Then, check out the project MLpseudoclient. This will create a local MLpseudoclient directory where the pseudo-client code will be located: 
              cvs co MLpseudoclient
              cd ./MLpseudoclient
The next thing would be to compile the client and server code. Both ./client and ./server directories have a compile.sh script that compile the client and server code. To run the pseudo-client server code, simply go to the ./scripts directory and run the script ./runMaster.sh. The script ./run_client.sh in the same directory will run the pseudo-client client code.

Stratos Efstathiadis - page was last modified