Patch proposal for log_duration
От | Guillaume Smet |
---|---|
Тема | Patch proposal for log_duration |
Дата | |
Msg-id | 1d4e0c10603300622o778da2adv66cc26d9cd7dc861@mail.gmail.com обсуждение исходный текст |
Ответы |
Re: Patch proposal for log_duration
|
Список | pgsql-patches |
Hi, This patch is a followup of a discussion on -hackers about the new behaviour of log_duration in 8.x ( http://archives.postgresql.org/pgsql-hackers/2006-03/msg00687.php ). As explained in the previous thread, we used to play with log_duration and log_min_error_statement to have the following behaviour: - log every duration so that we can have a global overview of the database activity; - log statement only for slowest queries. This setting is not yet possible with 8.x as log_duration only logs duration for queries logged with log_statement. On our production server, we cannot set log_statement to 'all' as it generates far too many writes on disks when logging the text of all the queries executed. But it's really useful for us to have a global overview of the database activity, not only slowest queries. FYI, our log file is used to generate this sort of report: http://pgfouine.projects.postgresql.org/reports/sample_hourly.html . This patch introduces back the ability to log every duration while keeping the new 8.x behaviour. You can set: - log_duration to 'none' = old off value - log_duration to 'logged' = on value for 8.x: log duration only if the query is logged by log_statement - log_duration to 'all' = on value for 7.4: we log duration for every statement executed. Patch is against current CVS head and make check is OK. I tested it with CVS and I backported it to 8.1.3 too to test it on one of our test server using 8.1. Please let me know if this patch needs any change as it's my first attempt to patch PostgreSQL. Regards, -- Guillaume
Вложения
В списке pgsql-patches по дате отправления: