)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"ae220cfa74e0f529a6b8cf35071353fa8296e033","unresolved":true,"context_lines":[{"line_number":7,"context_line":"Add a fake sys filesystem for the next patch"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"As we\u0027ll add a fixture for using sysfs, let\u0027s just accept to add a lot of"},{"line_number":10,"context_line":"sys files for Nova here :-)"},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"Partially-Implements: blueprint libvirt-cpu-state-mgmt"},{"line_number":13,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":1,"id":"92af2e7c_343f8e22","line":10,"updated":"2023-01-24 00:58:19.000000000","message":"so this is a subset of the sysfs form my home server.\n\ni dont object to including this as it was nice to be able to use this on my laptop when pocing arbiter but i just wanted to raise 2 points\n\nfirst this is more of sysfs then strictly needed for the current nova usecase.\nso you could cut this down.\n\nsecond the alternitive approch for arbiter that i considered was generating a subset of the filesystem dynmaically.\n\nwe could take that approch here if we wanted too.\n\nthese files where collected from a system with 2 12 core cpus with hyperthreading enebled\n\nso this models a phsyical host with 48 logical thread grop into 2 cpu sockets where each socket is a sperate numa node.\n\nit also has 256 GB of ram split evenly between both numa nodes.\n\nif i recall correctlly all cpu cores were online when this was generated too.\n\nfor thsoe interested this data i belive orginally came form https://github.com/SeanMooney/arbiterd/commit/20175d61f5cfb6ce96665f73724e4e7bcb57ebf0\n\nit was basically done by running \n\n“sudo cp -r --copy-contents -x --parents -P /sys/bus/node/ .“\n\ni belive i did that several time for differnt subpath\n\nso i would have rant it on /sys/bus/cpu for that sub path and the same for memory ectra.\n\ni guess my point is if we think this is too much to add as is we could trim it down or i can regenerate a more targeted set of data for the parts we actully use.","commit_id":"0237d6205faad2733b258bbeee19302a9d11edf2"}]}