Requirements

• python 2.4 or later
• rrdTool
• MySQL
• other python modules:

a) matplotlib

b) pyrrdTool

c) numpy

d) pysnmp

e) pysqlite

Modules

Currently, existing modules are:

• storeCons.py : this scrip has as a function to get the values of power consumption of the data center's machines in real-time via SNMP (Simple Network Management Protocol) and to store them in the database .

NB: A cron job for this script is lunched every 5 minutes

• getLoadFromRRD.py : this is the python script for moving average load values from rrd databases to the PowerMon SQL database

NB: A cron job for this script is lunched every hour

• updateDb.py : update the database¹ with the two configuration files: f_config_machines.cfg and f_config_pdus.cfg :

performs adding new pdu or a new machine to the database, and modifying their characteristics.

NB: The table 'PDU' must be modified before the table 'machine' because of foreign keys raisons

• viewCons.py The vizualisation script, that generates different plots via python plotting library MatplotLib,representing the power consumption and the average load of a machine during a certan period of time. This vizualisation tool is not finished yet

  1: PowerMon project supports all types of databases that are tolerated by sqlAlchemy (sqlite,mysql,posgresql,oracle...) 
       Therefore, we can specify which database we want to connect to , in the default section of f_config_pdus.cfg 
  

Classes

1. DbConn Objects: A DbConn object is responsible of database connection management and all the interactions with it

Constructor:

• class DbConn(dbName,dbType,username,password,host,port)

dbName,dbType,username,password,host,port are string arguments specified in the configuration file : f_config_pdu.cfg on section default

In PowerMon V1, "dbType" is either 'sqlite' or 'mysql' : other types may also be supported provided that they are supported by sqlAlchemy and that this constructor is appropriately modified

All class methods:

• create_tables() : Have to be called once, to create Powermon tables corresponding to the the following database schema :

• updatePDUtable(fConfigPDU): Collects changes(add or modify existing pdu) from config file f_config_pdus.cfg and updates the table "PDU"

• updateMachineTable(fConfigMachines): Does the same job as updatePDUtable() for table "machine"

• getIdOutlet(nomPDU,numOutlet): Returns outlet ID of the outlet number numOutlet belonging to pdu nomPDU

• viewDb(): Prints the contents of tables : Outlet, machine and PDU

• storeCons(nomMachine,cons,date): Inserts a new line in the table powerCons

• viewTableMachine(): Display only the table : machine

• viewMachineCons(nomMachine): Prints the power consumption values of the machine "nomMachine" whith the corresponding dates

• getPDUs(): Returns a list of python dictionaries , representing the PDUs registred in the database

• getNomMachine(numOutlet,IdPDU): Get the machine pluged-in on the outlet number "numOutlet" of the PDU whose Id is "IdPDU"

• getMachines_Names(): Returns a list of all machines' names registred in the database

• getPduMachines(pduName): Returns a set of machines' names correspondint to the connected machines to PDU "pduName"

• getCons(nomMachine,start_date,end_date): Returns a set of machine "nomMachine"'s power consumption values from "start_date" to "end_date" and another set of the corresponding dates

• getIdMachine(machineName): Returns the Id of the machine named "machineName"

• getLastLoadDate(IdMachine): Returns the most recent date in table "cpu_charge" where the machine Id is "IdMachine"

• storeLoad(IdMachine,dateList,avg1mn_list,avg5mn_list,avg15mn_list): Inserts a new line to the table cpu_charge whith the given entry values

• viewMachineLoad(nomMachine): Prints the load values of the machine "nomMachine" whith the corresponding dates

2. PDU Objects: A PDU object allows connexion to a supervisable pdu and extraction of machines' power consumption data

Constructor:

• class PDU()