Workernode-v0.0.2

From T2B Wiki
Revision as of 12:29, 26 August 2015 by Maintenance script (talk | contribs) (Created page with " === Basic OS Installation === The basic requirement for this benchmark is to have SLC 4.5 for the x86_64 architecture installed and running on the hardware. You may use ...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search


Basic OS Installation

The basic requirement for this benchmark is to have SLC 4.5 for the x86_64 architecture installed and running on the hardware. You may use PXE or CD installation from:

http://linuxsoft.cern.ch/scientific/45/x86_64/images/SL/pxeboot/

http://linuxsoft.cern.ch/scientific/45/x86_64/images/SL/boot.iso

Here is a summary of what the configuration must be:

  • take the "Personal Desktop" installation
  • let the installation procedure "Automatically partion" the disk(s)
  • keep the default firewall settings
  • put SElinux mode active
  • "Install default packages"
  • add a username and call it bench
  • at the CERN customization step select:
    • "No, I will do setup myself"
    • "Enable checks for updates"


Installing the Benchmarks

This installation will download around 2 GB of software.

  • Login as user bench
  • Run
wget http://mon.iihe.ac.be/trac/t2b/attachment/wiki/Workernode-v0.0.2/bench_1.tar?format=raw -O bench_1.tar; tar -xvf bench_1.tar;./install.sh


Power Consumption Benchmark

This benchmark takes between 10 to 15 minuntes.

  • Setup a measurement device between the power supply and the power plug
  • Run as bench:
./start_all.sh power
  • Collect the measured maximum power consumed before the start (in idle state) and the measured maximum power consumed during the benchmark run.


CPU Benchmarks

These benchmarks take between 12 to 24 hours and requires no manual intervention.

  • Run as bench:
./start_all.sh all
  • To check that the benchmarks are finished just run:
./start_all.sh check
  • When the benchmarks are finished, run the following command and follow the instructions:
./get_output.py


Useful Commands

You can also run :
  • ./start_all.sh stop to stop the benchmarks
  • ./start_all.sh check to check if the benchmarks are still running



Template:TracNotice