Личный кабинет
Укажите e-mail, на который будет выслан код восстановления пароля.
На указанный вами адрес e-mail был выслан код подтверждения аккаунта. Введите полученный код для продолжения:
Введите новый пароль два раза:
Hey,I upgraded my pg9.6 cluster to pg11.2.As it seems after the upgrade the duration of the same flow in my application raised from 13 minutes to 19 minutes. The test I did : 1.reset pg_stat_statements2.run the applicative flow3.collect everything from pg_stat_statementsI did this test on the env before the upgrade and after the upgrade. I took the sum of the total_time in pg_stat_statements and sumed it up.
my env settings : 60GB RAM16CPUregular HDpostgresql.conf settings : max_wal_size = 2GBmin_wal_size = 1GBwal_buffers = 16MBcheckpoint_completion_target = 0.9checkpoint_timeout = 30minstandard_conforming_strings = offmax_locks_per_transaction = 5000max_connections = 500random_page_cost = 4deadlock_timeout = 5sshared_preload_libraries = 'pg_stat_statements'track_activity_query_size = 32764log_directory = 'pg_log'enable_partitionwise_join = onenable_partitionwise_aggregate = onmax_worker_processes = 16 # (change requires restart)max_parallel_maintenance_workers = 8 # taken from max_parallel_workersmax_parallel_workers_per_gather = 8 # taken from max_parallel_workersmax_parallel_workers = 16maintenance_work_mem = 333MBwork_mem = 60MBshared_buffers = 15129MBeffective_cache_size = 30259MBThe conf file was used in 9.6 (without all the new parallel settings).Now the same queries run in both tests because it was the same flow. I will be happy to hear if u have any ideas without involving any queries changes..
В списке pgsql-performance по дате отправления: