)]}'
{"nova/compute/resource_tracker.py":[{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"422774fd1cd59531bb2df849dc50504a09ce235d","unresolved":false,"context_lines":[{"line_number":537,"context_line":"                # NOTE(tdurakov) this situation is possible for resize/cold"},{"line_number":538,"context_line":"                # migration when migration is finished but haven\u0027t yet"},{"line_number":539,"context_line":"                # confirmed/reverted in that case instance already changed host"},{"line_number":540,"context_line":"                # to destination and no matching happens"},{"line_number":541,"context_line":"                LOG.debug(\u0027Migration for instance %(uuid)s refers to \u0027"},{"line_number":542,"context_line":"                              \u0027another host\\\u0027s instance!\u0027,"},{"line_number":543,"context_line":"                          {\u0027uuid\u0027: migration.instance_uuid})"}],"source_content_type":"text/x-python","patch_set":1,"id":"1a95cdbc_73f5c8c9","line":540,"updated":"2016-10-05 15:48:48.000000000","message":"Is it possible to detect that this situation is currently happening and if so, log the debug message, otherwise, log the error message when it\u0027s a failure? I worry that by changing this to debug level that we may mask real error scenarios?","commit_id":"f085fbd7d3bfdf016a37ccc7e6e28786425f0e4e"},{"author":{"_account_id":5441,"name":"Andrew Laski","email":"andrew@lascii.com","username":"alaski"},"change_message_id":"fc07bc5af6cf9badca41d413960dd4afeb39cf2c","unresolved":false,"context_lines":[{"line_number":537,"context_line":"                # NOTE(tdurakov) this situation is possible for resize/cold"},{"line_number":538,"context_line":"                # migration when migration is finished but haven\u0027t yet"},{"line_number":539,"context_line":"                # confirmed/reverted in that case instance already changed host"},{"line_number":540,"context_line":"                # to destination and no matching happens"},{"line_number":541,"context_line":"                LOG.debug(\u0027Migration for instance %(uuid)s refers to \u0027"},{"line_number":542,"context_line":"                              \u0027another host\\\u0027s instance!\u0027,"},{"line_number":543,"context_line":"                          {\u0027uuid\u0027: migration.instance_uuid})"}],"source_content_type":"text/x-python","patch_set":1,"id":"1a95cdbc_97b757f4","line":540,"in_reply_to":"1a95cdbc_73f5c8c9","updated":"2016-10-06 21:23:13.000000000","message":"I\u0027m not sure that this happens in a real failure situation, though of course I could be wrong.\n\nThe tricky thing here is that an instance lives in one place but a migration refers to two. So while migrating from A to B the update_available_resource periodic task on both hosts is going to see the migration and one of them will always fail to find an associated instance. So we can expect this logging to occur fairly regularly.\n\nIn order to log this as an error the migration object will need some way to say definitively that the instance should exist on this host and then check that when the instance is not found. Perhaps that can be derived from a combination of migration.status and source or dest compute fields, but I\u0027m not familiar enough with the object to say.\n\nAll that is to say that I think the change to debug is fine and we need a better way to determine errors when pairing migrations to instances.\n\n\nAlternatively instead of grabbing all migrations for this host and then comparing to instances on this host the periodic job could just grab migrations for instances on this host and work off of that. There may be some reason for not doing that that I\u0027m overlooking though.","commit_id":"f085fbd7d3bfdf016a37ccc7e6e28786425f0e4e"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"2cdb4330f9f02425fc1129b2547840ae409e9613","unresolved":false,"context_lines":[{"line_number":537,"context_line":"                # NOTE(tdurakov) this situation is possible for resize/cold"},{"line_number":538,"context_line":"                # migration when migration is finished but haven\u0027t yet"},{"line_number":539,"context_line":"                # confirmed/reverted in that case instance already changed host"},{"line_number":540,"context_line":"                # to destination and no matching happens"},{"line_number":541,"context_line":"                LOG.debug(\u0027Migration for instance %(uuid)s refers to \u0027"},{"line_number":542,"context_line":"                              \u0027another host\\\u0027s instance!\u0027,"},{"line_number":543,"context_line":"                          {\u0027uuid\u0027: migration.instance_uuid})"}],"source_content_type":"text/x-python","patch_set":1,"id":"1a95cdbc_bb641eb9","line":540,"in_reply_to":"1a95cdbc_97b757f4","updated":"2016-10-07 14:58:02.000000000","message":"Yeah, I see your point Andrew. This is, coincidentally, why it would be better to have these kinds of log statements in a conductor thread that understands both sides of the migration operation :)","commit_id":"f085fbd7d3bfdf016a37ccc7e6e28786425f0e4e"},{"author":{"_account_id":16907,"name":"Diana Clarke","email":"diana.joan.clarke@gmail.com","username":"diana-clarke"},"change_message_id":"4a0e890f1fdc9a06b5859a448f21d6a53b8d7440","unresolved":false,"context_lines":[{"line_number":539,"context_line":"                # confirmed/reverted in that case instance already changed host"},{"line_number":540,"context_line":"                # to destination and no matching happens"},{"line_number":541,"context_line":"                LOG.debug(\u0027Migration for instance %(uuid)s refers to \u0027"},{"line_number":542,"context_line":"                              \u0027another host\\\u0027s instance!\u0027,"},{"line_number":543,"context_line":"                          {\u0027uuid\u0027: migration.instance_uuid})"},{"line_number":544,"context_line":""},{"line_number":545,"context_line":"    @utils.synchronized(COMPUTE_RESOURCE_SEMAPHORE)"}],"source_content_type":"text/x-python","patch_set":1,"id":"fa6399be_8ad3b0d4","line":542,"range":{"start_line":542,"start_character":31,"end_line":542,"end_character":56},"updated":"2016-10-20 11:48:10.000000000","message":"Thoughts on a less alarming log message if this isn\u0027t necessarily an error case? Like removing the exclamation mark, etc... I\u0027m asking because operators are asking.","commit_id":"f085fbd7d3bfdf016a37ccc7e6e28786425f0e4e"}]}