)]}'
{"swift/container/backend.py":[{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"d3c6df71cc7a3e5f24d326df7318d2bce63e3963","unresolved":false,"context_lines":[{"line_number":1802,"context_line":"        lockpath \u003d \u0027%s/.sharding\u0027 % self.db_dir"},{"line_number":1803,"context_line":"        try:"},{"line_number":1804,"context_line":"            fd \u003d os.open(lockpath, os.O_WRONLY | os.O_CREAT)"},{"line_number":1805,"context_line":"            yield fd"},{"line_number":1806,"context_line":"        finally:"},{"line_number":1807,"context_line":"            os.close(fd)"},{"line_number":1808,"context_line":"            os.unlink(lockpath)"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_a352bc11","side":"PARENT","line":1805,"updated":"2018-04-25 18:58:45.000000000","message":"Especially concerning was the fact that we never called fcntl.flock -- I don\u0027t think this was *ever* doing what we wanted.","commit_id":"2bc60b649bb2bc2883e6a361be1200094cae1678"},{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"ce8613add74b218c1331414d241a68bc247d22bc","unresolved":false,"context_lines":[{"line_number":1802,"context_line":"        lockpath \u003d \u0027%s/.sharding\u0027 % self.db_dir"},{"line_number":1803,"context_line":"        try:"},{"line_number":1804,"context_line":"            fd \u003d os.open(lockpath, os.O_WRONLY | os.O_CREAT)"},{"line_number":1805,"context_line":"            yield fd"},{"line_number":1806,"context_line":"        finally:"},{"line_number":1807,"context_line":"            os.close(fd)"},{"line_number":1808,"context_line":"            os.unlink(lockpath)"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_0cf635d8","side":"PARENT","line":1805,"in_reply_to":"9f6a8fd7_a352bc11","updated":"2018-04-25 21:06:30.000000000","message":"+1 it either needs fixing or removing :)","commit_id":"2bc60b649bb2bc2883e6a361be1200094cae1678"}],"swift/container/sharder.py":[{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"d3c6df71cc7a3e5f24d326df7318d2bce63e3963","unresolved":false,"context_lines":[{"line_number":984,"context_line":"            for objects, info in self.yield_objects("},{"line_number":985,"context_line":"                    source_broker, shard_range, policy_index,"},{"line_number":986,"context_line":"                    since_row\u003dcleaving_context.last_cleave_to_row):"},{"line_number":987,"context_line":"                shard_broker.merge_items(objects)"},{"line_number":988,"context_line":""},{"line_number":989,"context_line":"        own_shard_range \u003d broker.get_own_shard_range()"},{"line_number":990,"context_line":"        if shard_range.includes(own_shard_range):"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_8339b8a4","line":987,"updated":"2018-04-25 18:58:45.000000000","message":"I do wonder a bit if we still need to worry about the replicator coming along when the shard we\u0027re cleaving out is only partially filled... but what we had before certainly wasn\u0027t guarding against that.","commit_id":"d46102a6a45c5e454c1121594c016a0e2a118297"},{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"004a8a9de65c9d7badb15a1021e0608db1e1b248","unresolved":false,"context_lines":[{"line_number":984,"context_line":"            for objects, info in self.yield_objects("},{"line_number":985,"context_line":"                    source_broker, shard_range, policy_index,"},{"line_number":986,"context_line":"                    since_row\u003dcleaving_context.last_cleave_to_row):"},{"line_number":987,"context_line":"                shard_broker.merge_items(objects)"},{"line_number":988,"context_line":""},{"line_number":989,"context_line":"        own_shard_range \u003d broker.get_own_shard_range()"},{"line_number":990,"context_line":"        if shard_range.includes(own_shard_range):"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_394a0727","line":987,"in_reply_to":"9f6a8fd7_6c07b1cb","updated":"2018-05-10 13:57:43.000000000","message":"I checked on this - we do merge a shards own range when first getting the broker so no danger of the shard db getting objects merged but no getting its shard range","commit_id":"d46102a6a45c5e454c1121594c016a0e2a118297"},{"author":{"_account_id":7847,"name":"Alistair Coles","email":"alistairncoles@gmail.com","username":"acoles"},"change_message_id":"ce8613add74b218c1331414d241a68bc247d22bc","unresolved":false,"context_lines":[{"line_number":984,"context_line":"            for objects, info in self.yield_objects("},{"line_number":985,"context_line":"                    source_broker, shard_range, policy_index,"},{"line_number":986,"context_line":"                    since_row\u003dcleaving_context.last_cleave_to_row):"},{"line_number":987,"context_line":"                shard_broker.merge_items(objects)"},{"line_number":988,"context_line":""},{"line_number":989,"context_line":"        own_shard_range \u003d broker.get_own_shard_range()"},{"line_number":990,"context_line":"        if shard_range.includes(own_shard_range):"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_6c07b1cb","line":987,"in_reply_to":"9f6a8fd7_8339b8a4","updated":"2018-04-25 21:06:30.000000000","message":"yeah I want to move merging shard ranges ahead of merging objects. I think that is a safer order if only some get replicated.","commit_id":"d46102a6a45c5e454c1121594c016a0e2a118297"},{"author":{"_account_id":1179,"name":"Clay Gerrard","email":"clay.gerrard@gmail.com","username":"clay-gerrard"},"change_message_id":"f0469e88d773eb8f5f6fc523968528477408d211","unresolved":false,"context_lines":[{"line_number":805,"context_line":""},{"line_number":806,"context_line":"        if broker.get_info()[\u0027id\u0027] !\u003d info[\u0027id\u0027]:"},{"line_number":807,"context_line":"            # the db changed - don\u0027t remove any objects"},{"line_number":808,"context_line":"            success \u003d False"},{"line_number":809,"context_line":"        else:"},{"line_number":810,"context_line":"            # remove objects up to the max row of the db sampled prior to"},{"line_number":811,"context_line":"            # the first object yielded for this destination; objects added"}],"source_content_type":"text/x-python","patch_set":2,"id":"9f6a8fd7_d930aba1","line":808,"updated":"2018-04-30 20:34:21.000000000","message":"this is the optimistic lock strategy - only cleanup if the database hasn\u0027t changed - LGTM\n\nunless we took this lock every time we added rows taking it when removing doesn\u0027t help","commit_id":"d9be8caae2d43159bd476da979ce7981c38efd1c"}]}
