BenchMarks < HEPrc

Tags: view all tags
---+ VM Benchmarks

CERN documentation http://bmkwg.web.cern.ch/bmkwg/docs/HowToRun.html

Installation:
   * bellecs.heprc.uvic.ca in /root/HEPbm.yaml
   * condor.heprc.uvic.ca in 

Run with CVMFS
<verbatim style="background: #CCFFFF">
export BMK_ROOTDIR=/cvmfs/atlas.cern.ch/repo/benchmarks/cern/current
cern-benchmark --benchmarks="fastBmk;whetstone" --cloud=uvic --queue_host=dashb-test-mb.cern.ch
--queue_port=61113 --username=xxx --password=`cat /pwdfile` --topic=/topic/vm.spec
</verbatim>


Kibana web page:  https://dashb-es-dev.cern.ch:444/#/dashboard/sdccloud-int


To extract the information in a JSON format (log into lxplus)
<verbatim style="background: #CCFFFF">
cern-get-sso-cookie --krb -r -u https://dashb-es-dev.cern.ch -o ~/private/ssocookie_es.txt
curl -L --cookie ~/private/ssocookie_es.txt --cookie-jar ~/private/ssocookie_es.txt  -XGET 'https://dashb-es-dev.cern.ch:443/sdccloud-int/_search?pretty' -d "$(cat bmk_es_query)”
</verbatim>

The query file is found in QueryFile

One can reformat the JSON output file into a more readable format by pipping the output to  jq "."

Then extract the required information (on macsob2 or lxplus); written to out.list.csv
<verbatim style="background: #CCFFFF">
python SimpleEStoCSV.py -f out.list
</verbatim>

Then delete the first line and replace the commas with spaces
<verbatim style="background: #CCFFFF">
 sed 's/,/ /g' out.list.csv >out.list.csv.2
</verbatim>

Now it is ready for ROOT
Topic revision: r2 - 2016-07-19 - rsobie