)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":8788,"name":"Miguel Angel Ajo","email":"mangelajo@redhat.com","username":"mangelajo"},"change_message_id":"f2eb282bde74fa6376267ba6e45c280cd0bb165c","unresolved":false,"context_lines":[{"line_number":23,"context_line":"to work."},{"line_number":24,"context_line":""},{"line_number":25,"context_line":"We can make test modifications to solve these issues. The alternative"},{"line_number":26,"context_line":"which I\u0027m proposing is to make keepalived.spawn blocking."},{"line_number":27,"context_line":""},{"line_number":28,"context_line":"Change-Id: Ie77d406eaaf7f77edd4f598947999be4adf3d249"},{"line_number":29,"context_line":"Closes-Bug: #1567023"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"9a061dce_ad1e7eb8","line":26,"updated":"2016-04-07 08:23:35.000000000","message":"I\u0027d avoid making it blocking, because then spawning lots of routers would be slower and we would stop relying on the \"external parallelism\".\n\nI believe it\u0027s reasonable to wait_until_true here in tests.","commit_id":"b80bc2d50ddcdf7a0d560c259c1c2e9c73367064"}],"neutron/agent/linux/keepalived.py":[{"author":{"_account_id":8788,"name":"Miguel Angel Ajo","email":"mangelajo@redhat.com","username":"mangelajo"},"change_message_id":"f2eb282bde74fa6376267ba6e45c280cd0bb165c","unresolved":false,"context_lines":[{"line_number":405,"context_line":"                                      service_name\u003dKEEPALIVED_SERVICE_NAME,"},{"line_number":406,"context_line":"                                      monitored_process\u003dkeepalived_pm)"},{"line_number":407,"context_line":""},{"line_number":408,"context_line":"        utils.wait_until_true("},{"line_number":409,"context_line":"            lambda: keepalived_pm.active,"},{"line_number":410,"context_line":"            timeout\u003d10,"},{"line_number":411,"context_line":"            sleep\u003d0.01,"}],"source_content_type":"text/x-python","patch_set":2,"id":"9a061dce_48123058","line":408,"updated":"2016-04-07 08:23:35.000000000","message":"Do we find any other reason (other than testing) to wait for the process to be active?.\n\nBy doing this we kill any parallelization, and make the l3-agent startup time much slower when you have hundreds of routers.\n\nIf it does not happen process_monitor will start complaining.","commit_id":"b80bc2d50ddcdf7a0d560c259c1c2e9c73367064"}]}