Re: Add last_vacuum_index_scans in pg_stat_all_tables
От | Fujii Masao |
---|---|
Тема | Re: Add last_vacuum_index_scans in pg_stat_all_tables |
Дата | |
Msg-id | bc26e96b-181f-060d-b13c-ff5588d6d386@oss.nttdata.com обсуждение исходный текст |
Ответ на | Re: Add last_vacuum_index_scans in pg_stat_all_tables (Ken Kato <katouknl@oss.nttdata.com>) |
Ответы |
Re: Add last_vacuum_index_scans in pg_stat_all_tables
|
Список | pgsql-hackers |
On 2022/09/16 13:23, Ken Kato wrote: >> Regression is failing on all platforms; please correct that and >> resubmit the patch. > > Hi, > > Thank you for the review! > I fixed it and resubmitting the patch. Could you tell me why the number of index scans should be tracked for each table? Instead, isn't it enough to have one global counter, to check whether the current setting of maintenance_work_mem is sufficient or not? That is, I'm thinking to have something like pg_stat_vacuum view that reports, for example, the number of vacuum runs, the total number of index scans, the maximum number of index scans by one vacuum run, the number of cancellation of vacuum because of lock conflicts, etc. If so, when these global counters are high or increasing, we can think that it may worth tuning maintenance_work_mem. Regards, -- Fujii Masao Advanced Computing Technology Center Research and Development Headquarters NTT DATA CORPORATION
В списке pgsql-hackers по дате отправления: