)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"dcebc0ca348eecbdcd1e5bd29e7ebf30383d64d6","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"* Progress_watermark is design for progress_timeout, and progress_timeout"},{"line_number":8,"context_line":"  is design for stuck of libvirt no matter progress_watermark is lesser than"},{"line_number":9,"context_line":"  data_remaining or greater than data_remaining."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"* When progress_watermark \u003e data_remaining, it\u0027s meat that libvirtd has begin LM"},{"line_number":12,"context_line":""}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"ba5da102_46b5ba7b","line":9,"updated":"2016-11-03 14:54:39.000000000","message":"THis is not a properly structured commit message - you must have a single line summary followed by blank line","commit_id":"01806972e918b47a2b3f61a5f6695161ed728629"}],"nova/virt/libvirt/driver.py":[{"author":{"_account_id":12299,"name":"Paweł Koniszewski","email":"pawel.koniszewski@gmail.com","username":"pawel.koniszewski"},"change_message_id":"61e8f659df894d06eaf70bff695f5ec31aed7b1b","unresolved":false,"context_lines":[{"line_number":6126,"context_line":"                elapsed \u003d now - start"},{"line_number":6127,"context_line":""},{"line_number":6128,"context_line":"                if ((progress_watermark \u003d\u003d 0) or"},{"line_number":6129,"context_line":"                    (progress_watermark !\u003d info.data_remaining)):"},{"line_number":6130,"context_line":"                    progress_watermark \u003d info.data_remaining"},{"line_number":6131,"context_line":"                    progress_time \u003d now"},{"line_number":6132,"context_line":""}],"source_content_type":"text/x-python","patch_set":1,"id":"5a8bc5a2_779330fb","line":6129,"range":{"start_line":6129,"start_character":21,"end_line":6129,"end_character":62},"updated":"2016-09-22 08:00:40.000000000","message":"Given how dynamic LM process is and that libvirt reports data in bytes, this will be different in every iteration, so that this condition will always be True. We can\u0027t update progress_watermark and progress_time this way. Also please refer to my comment in previous patch set.","commit_id":"1392fcb90f6bf738e5f370c9bd6ee86b0c01fb9a"},{"author":{"_account_id":12299,"name":"Paweł Koniszewski","email":"pawel.koniszewski@gmail.com","username":"pawel.koniszewski"},"change_message_id":"50f60d3a976bd3dee158633908f422260556e58b","unresolved":false,"context_lines":[{"line_number":6126,"context_line":"                elapsed \u003d now - start"},{"line_number":6127,"context_line":""},{"line_number":6128,"context_line":"                if ((progress_watermark \u003d\u003d 0) or"},{"line_number":6129,"context_line":"                    (progress_watermark !\u003d info.data_remaining)):"},{"line_number":6130,"context_line":"                    progress_watermark \u003d info.data_remaining"},{"line_number":6131,"context_line":"                    progress_time \u003d now"},{"line_number":6132,"context_line":""}],"source_content_type":"text/x-python","patch_set":2,"id":"5a8bc5a2_5c23b405","line":6129,"range":{"start_line":6129,"start_character":21,"end_line":6129,"end_character":62},"updated":"2016-09-22 09:12:33.000000000","message":"This condition is always true, so every time when nova gets data from libvirt progress_watermark will be different than info.data_remaining. This is due to:\n* QEMU at the beginning of each iteration calculates data to be transferred in this particular iteration using formula data_remaining \u003d number_of_drity_pages * size_of_a_single_page\n* During single iteration data_remaining goes down to 0 and then is refreshed at the beginning of subsequent iteration. \n* data_remaining is reported in bytes so given how dynamic live migration process is it will be different every time nova receives data from libvirt. I doubt we will ever hit the same value.\nWe can\u0027t say that when those values are different there was a progress in live migration. Also we really can\u0027t base on info.data_remaining to determine whether there is really any progress in LM.","commit_id":"01806972e918b47a2b3f61a5f6695161ed728629"}]}