)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"0427c75e_aff416e6","updated":"2022-10-20 20:30:19.000000000","message":"That\u0027s a really good starting point for mysql/galera config reference IMO.\nExcept the cleanupjobs nit, I am not against merging that as a first shot.","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":308,"name":"Thierry Carrez","email":"thierry@openstack.org","username":"ttx"},"change_message_id":"762955ad027144cfac9f8076040a9676549c897f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"56e0230b_65e57a4e","updated":"2022-10-21 10:29:42.000000000","message":"Looks great!","commit_id":"d09670a9c7f63a954099100b58b20cad63053d19"},{"author":{"_account_id":13252,"name":"Dr. Jens Harbott","display_name":"Jens Harbott (frickler)","email":"frickler@offenerstapel.de","username":"jrosenboom"},"change_message_id":"7935f8c50894497c8becf964941d057b7589aabd","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"96882f17_5bfb9212","updated":"2022-10-21 09:36:45.000000000","message":"Only +1 from me in order not to self-approve my updates.","commit_id":"d09670a9c7f63a954099100b58b20cad63053d19"}],"doc/source/other/database.rst":[{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":51,"context_line":""},{"line_number":52,"context_line":"  [mysqld]"},{"line_number":53,"context_line":"  max_connections\u003d4000"},{"line_number":54,"context_line":"  max_statement_time\u003d20"},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"In order to ensure your cluster runs smoothly we recommend you limit the amount of connections and the time statements can be executed."},{"line_number":57,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"446583ed_aa4d7a1f","line":54,"range":{"start_line":54,"start_character":2,"end_line":54,"end_character":20},"updated":"2022-10-20 20:30:19.000000000","message":"I am not entirely sure that all our requests are running under 20sec.\nWe used to have long running requests when the DB was big.\nSince we used osarchiver to cleanup the DB, we dont have much, but 20sec seems a low value to me.\nI will do some checks on our infras.","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":29074,"name":"Felix Huettner","email":"felix.huettner@digits.schwarz","username":"felix.huettner"},"change_message_id":"76b6719780f2ab67f4cf46e2215f2eba5b4f1913","unresolved":true,"context_lines":[{"line_number":51,"context_line":""},{"line_number":52,"context_line":"  [mysqld]"},{"line_number":53,"context_line":"  max_connections\u003d4000"},{"line_number":54,"context_line":"  max_statement_time\u003d20"},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"In order to ensure your cluster runs smoothly we recommend you limit the amount of connections and the time statements can be executed."},{"line_number":57,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"68a18b6b_a363b395","line":54,"range":{"start_line":54,"start_character":2,"end_line":54,"end_character":20},"in_reply_to":"446583ed_aa4d7a1f","updated":"2022-10-21 06:29:19.000000000","message":"Besides of our initial cleanup jobs we never had the issue of exceeding 20 sec.\n\nBut the database cleanup is a good point. I added a section for that.","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":58,"context_line":"The value for `max_connections` should be set based on actual tests (testing with a lot of idle connections is fine)."},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"The value `20` seconds for `max_statement_time` is enough for all normal use-cases we know of."},{"line_number":61,"context_line":"You might only run into issues with regular nova cleanupjobs if they do not run often enough."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Replication stability"},{"line_number":64,"context_line":"^^^^^^^^^^^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"468e2c63_51d44a22","line":61,"range":{"start_line":61,"start_character":49,"end_line":61,"end_character":60},"updated":"2022-10-20 20:30:19.000000000","message":"cleanup jobs","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":58,"context_line":"The value for `max_connections` should be set based on actual tests (testing with a lot of idle connections is fine)."},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"The value `20` seconds for `max_statement_time` is enough for all normal use-cases we know of."},{"line_number":61,"context_line":"You might only run into issues with regular nova cleanupjobs if they do not run often enough."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Replication stability"},{"line_number":64,"context_line":"^^^^^^^^^^^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"6ed47db9_9ca1e585","line":61,"updated":"2022-10-20 20:30:19.000000000","message":"what happen then if the request take more than 20sec, is nova never able to cleanup?","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":29074,"name":"Felix Huettner","email":"felix.huettner@digits.schwarz","username":"felix.huettner"},"change_message_id":"76b6719780f2ab67f4cf46e2215f2eba5b4f1913","unresolved":true,"context_lines":[{"line_number":58,"context_line":"The value for `max_connections` should be set based on actual tests (testing with a lot of idle connections is fine)."},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"The value `20` seconds for `max_statement_time` is enough for all normal use-cases we know of."},{"line_number":61,"context_line":"You might only run into issues with regular nova cleanupjobs if they do not run often enough."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Replication stability"},{"line_number":64,"context_line":"^^^^^^^^^^^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"53f9d651_ad2226bc","line":61,"range":{"start_line":61,"start_character":49,"end_line":61,"end_character":60},"in_reply_to":"468e2c63_51d44a22","updated":"2022-10-21 06:29:19.000000000","message":"fixed","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":29074,"name":"Felix Huettner","email":"felix.huettner@digits.schwarz","username":"felix.huettner"},"change_message_id":"76b6719780f2ab67f4cf46e2215f2eba5b4f1913","unresolved":true,"context_lines":[{"line_number":58,"context_line":"The value for `max_connections` should be set based on actual tests (testing with a lot of idle connections is fine)."},{"line_number":59,"context_line":""},{"line_number":60,"context_line":"The value `20` seconds for `max_statement_time` is enough for all normal use-cases we know of."},{"line_number":61,"context_line":"You might only run into issues with regular nova cleanupjobs if they do not run often enough."},{"line_number":62,"context_line":""},{"line_number":63,"context_line":"Replication stability"},{"line_number":64,"context_line":"^^^^^^^^^^^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"091817c6_c1d803a2","line":61,"in_reply_to":"6ed47db9_9ca1e585","updated":"2022-10-21 06:29:19.000000000","message":"also added a new section for that `--max-rows` helps here","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":73,"context_line":""},{"line_number":74,"context_line":"This setting sets the replication thread to realtime scheduling on the kernel side."},{"line_number":75,"context_line":"If you run galera as a non-privileged user (as you hopefully do) galera will need `CAP_SYS_NICE` to be allowed to change the priority."},{"line_number":76,"context_line":"If you run inside a container environment you might need to set `kernel.sched_rt_runtime_us\u003d-1` (allthough that is suboptimal."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Performance"},{"line_number":79,"context_line":"^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"312c438b_2d1e27c7","line":76,"updated":"2022-10-20 20:30:19.000000000","message":"this seems a great value for large-scale config guideline, thanks!","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":29074,"name":"Felix Huettner","email":"felix.huettner@digits.schwarz","username":"felix.huettner"},"change_message_id":"76b6719780f2ab67f4cf46e2215f2eba5b4f1913","unresolved":false,"context_lines":[{"line_number":73,"context_line":""},{"line_number":74,"context_line":"This setting sets the replication thread to realtime scheduling on the kernel side."},{"line_number":75,"context_line":"If you run galera as a non-privileged user (as you hopefully do) galera will need `CAP_SYS_NICE` to be allowed to change the priority."},{"line_number":76,"context_line":"If you run inside a container environment you might need to set `kernel.sched_rt_runtime_us\u003d-1` (allthough that is suboptimal."},{"line_number":77,"context_line":""},{"line_number":78,"context_line":"Performance"},{"line_number":79,"context_line":"^^^^^^^^^^^"}],"source_content_type":"text/x-rst","patch_set":1,"id":"b280cc55_387bd759","line":76,"in_reply_to":"312c438b_2d1e27c7","updated":"2022-10-21 06:29:19.000000000","message":"Ack","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":109,"context_line":""},{"line_number":110,"context_line":"Entering all servers with `backup` at the end ensures that haproxy will only choose the second server if the first is offline."},{"line_number":111,"context_line":""},{"line_number":112,"context_line":"You should note the `timeout client` setting here, as it is relevant to the openstack configuration."},{"line_number":113,"context_line":""},{"line_number":114,"context_line":"-----------------------"},{"line_number":115,"context_line":"Openstack Configuration"}],"source_content_type":"text/x-rst","patch_set":1,"id":"f1dd40c8_92fc8078","line":112,"updated":"2022-10-20 20:30:19.000000000","message":"We did something similar with a slight difference.\nSo we can keep your proposal here for a first shot and I will bring a patch on top of it to show you.","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"},{"author":{"_account_id":11583,"name":"Arnaud Morin","email":"arnaud.morin@gmail.com","username":"arnaudmorin"},"change_message_id":"217e236be5dca52156a91b8b76f9c95b51d2365a","unresolved":true,"context_lines":[{"line_number":128,"context_line":"This ensures connections are recreated on the openstack side first before the reverse proxy is forcing the connection to terminate."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"The `max_pool_size` and `max_overflow` define the amount of connections an individual thread is allowed to have."},{"line_number":131,"context_line":"You will need to set this on experience (allthough the above should be a good start)."},{"line_number":132,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"20b35263_fa13d789","line":131,"updated":"2022-10-20 20:30:19.000000000","message":"These are the values I am currently questioning myself about.\nWe have lower values on our end because we dont want to reach the max_connection set on mariadb side (we already did, and you dont want that to happen, trust me).\n\nI\u0027d like to see if someone else (maybe belmiro?) have a different way to setting this.\nThis is not yet clear to me.","commit_id":"02802ca941943692f580fa2a11725d10c8214c84"}]}