Data rushing in for the HTo4L analysis (Chad)
Chad asked for the Higgs to 4 lepton analysis for the Junior Lab Open data project to have samples copied to submit. I use this as a test for seeing the bandwidth we can copy at. FiBS* is processing 10 downloads in parallel on each of the submit50-59 servers (they carry the gluster drives). This means 100 parallel processes.
Performance is pretty good, little spikes. I think we are hitting the bottleneck.
$ du -sh /data/submit/cms/store/
897G /data/submit/cms/store/- FiBS at: https://github.com/cpausmit/FiBS