• Print
• View
• History
• Login

Install Monitoring

Low level monitoring

• Hardware level: IMPI?
• System level: SNMP+MRTG

High level monitoring

Detailed statistics of how many jobs are run, by whom, which VOs etc cannot be done with SNMP and MRTG. We have not yet chosen a monitoring system. Below are some options:

• http://ganglia.info/
• MoniKa at http://ka-tools.imag.fr/
• PBS tools http://www.osc.edu/~troy/pbs/
• Home-made Torque monitoring
• Stocking the Toolbox: Tools No Cluster Admin Should Be Without Linux Magazine article
• http://www-rcf.usc.edu/~garrick/perl-PBS/
  • pbstop seems nice
• http://homepages.inf.ed.ac.uk/s0239160/misc/lspbs/lspbs.html

One thing to consider is that for grid jobs, when they reach PBS they do not have user information, just VO.

Ongoing installs:

• Install Ganglia
• Install Panda