)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":26458,"name":"Brin Zhang","email":"zhangbailin@inspur.com","username":"zhangbailin"},"change_message_id":"f97027824574e04440e3f0907ac9410130ddca67","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"c24b1cf4_afd63f65","updated":"2024-07-29 01:12:51.000000000","message":"LGTM","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":31412,"name":"Wenping Song","email":"songwenping@inspur.com","username":"songwenping"},"change_message_id":"263ad13992cd65a660606f943cc55c30243ed67b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"ce832844_64ac0397","updated":"2024-07-29 10:09:29.000000000","message":"Thanks Arnaud Morin for review again.","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"ff9712766fc453feb7a783602da733d64d1b470b","unresolved":true,"context_lines":[],"source_content_type":"","patch_set":3,"id":"1491e515_4e91b462","updated":"2024-07-28 15:10:18.000000000","message":"That\u0027s a valuable story.\nThank you for your submission!\n\nI have few questions that I think could help people reading your story later.\nIf you dont mind answering to them by amending a little bit your commit?","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"}],"doc/source/stories/2024-06-10.rst":[{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"ff9712766fc453feb7a783602da733d64d1b470b","unresolved":true,"context_lines":[{"line_number":38,"context_line":"::"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"    [DEFAULT]"},{"line_number":41,"context_line":"    max_connections \u003d 100000"},{"line_number":42,"context_line":"    thread_cache_size \u003d 10000"},{"line_number":43,"context_line":""},{"line_number":44,"context_line":"2. Optimize the Rabbitmq configuration of message middleware"}],"source_content_type":"text/x-rst","patch_set":3,"id":"3d0f5634_be9d318e","line":41,"updated":"2024-07-28 15:10:18.000000000","message":"Can you maybe explain that this is a mariadb (or mysql?) parameter, not an openstack one?","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":31412,"name":"Wenping Song","email":"songwenping@inspur.com","username":"songwenping"},"change_message_id":"263ad13992cd65a660606f943cc55c30243ed67b","unresolved":true,"context_lines":[{"line_number":38,"context_line":"::"},{"line_number":39,"context_line":""},{"line_number":40,"context_line":"    [DEFAULT]"},{"line_number":41,"context_line":"    max_connections \u003d 100000"},{"line_number":42,"context_line":"    thread_cache_size \u003d 10000"},{"line_number":43,"context_line":""},{"line_number":44,"context_line":"2. Optimize the Rabbitmq configuration of message middleware"}],"source_content_type":"text/x-rst","patch_set":3,"id":"d9dbd3f1_212b715f","line":41,"in_reply_to":"3d0f5634_be9d318e","updated":"2024-07-29 10:09:29.000000000","message":"This is mariadb parameter, we use mariadb cluster.","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"ff9712766fc453feb7a783602da733d64d1b470b","unresolved":true,"context_lines":[{"line_number":51,"context_line":"::"},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"    [DEFAULT]"},{"line_number":54,"context_line":"    maximum\u003d20000"},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"The max number of conns for RabbitMQ is estimated through cacluating the"},{"line_number":57,"context_line":"conns of nova and cinder components. We have 3000 compute nodes, 3 control"}],"source_content_type":"text/x-rst","patch_set":3,"id":"841e6bbe_9320ec74","line":54,"updated":"2024-07-28 15:10:18.000000000","message":"Idem, this is a tuning that you set on rabbitmq side I suppose?\nWould you mind letting us know in which file you set this?","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":31412,"name":"Wenping Song","email":"songwenping@inspur.com","username":"songwenping"},"change_message_id":"263ad13992cd65a660606f943cc55c30243ed67b","unresolved":true,"context_lines":[{"line_number":51,"context_line":"::"},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"    [DEFAULT]"},{"line_number":54,"context_line":"    maximum\u003d20000"},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"The max number of conns for RabbitMQ is estimated through cacluating the"},{"line_number":57,"context_line":"conns of nova and cinder components. We have 3000 compute nodes, 3 control"}],"source_content_type":"text/x-rst","patch_set":3,"id":"104310de_fad74d37","line":54,"in_reply_to":"841e6bbe_9320ec74","updated":"2024-07-29 10:09:29.000000000","message":"This tuning is set on rabbitmq server side, so we set this in rabbitmq.conf.","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"ff9712766fc453feb7a783602da733d64d1b470b","unresolved":true,"context_lines":[{"line_number":65,"context_line":"service, the connection to Rabbitmq was not released, and other coroutines"},{"line_number":66,"context_line":"were unable to obtain the connection, resulting in message backlog."},{"line_number":67,"context_line":"We reduce the risk of message backlog by modifying the code, adjusting"},{"line_number":68,"context_line":"the message timeout mechanism, and increasing the timeout duration."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"4. Increase the number of allocation candidates returned from Placement"},{"line_number":71,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"dd8b1070_d19ac0a0","line":68,"updated":"2024-07-28 15:10:18.000000000","message":"I remember that nova is sending a lot of messages in periodic basics.\nMaybe you tuned the nova periodic task parameter as well?\nhttps://docs.openstack.org/operations-guide/ops-advanced-configuration.html","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"},{"author":{"_account_id":31412,"name":"Wenping Song","email":"songwenping@inspur.com","username":"songwenping"},"change_message_id":"263ad13992cd65a660606f943cc55c30243ed67b","unresolved":true,"context_lines":[{"line_number":65,"context_line":"service, the connection to Rabbitmq was not released, and other coroutines"},{"line_number":66,"context_line":"were unable to obtain the connection, resulting in message backlog."},{"line_number":67,"context_line":"We reduce the risk of message backlog by modifying the code, adjusting"},{"line_number":68,"context_line":"the message timeout mechanism, and increasing the timeout duration."},{"line_number":69,"context_line":""},{"line_number":70,"context_line":"4. Increase the number of allocation candidates returned from Placement"},{"line_number":71,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"86182e54_3746471a","line":68,"in_reply_to":"dd8b1070_d19ac0a0","updated":"2024-07-29 10:09:29.000000000","message":"The period task messages is smaller than build instance requests, eapecially in steady state, the rabbitmq can process these messages easily, so we barely tuned the nova periodic task parameter.","commit_id":"401969e8ceef40d8556680de89ffd894af483fa7"}]}