Re: Non-pausing table scan on 9.6 replica?
От | Andreas Kretschmer |
---|---|
Тема | Re: Non-pausing table scan on 9.6 replica? |
Дата | |
Msg-id | af14a041-eb0d-05f2-8d84-82f6579df168@a-kretschmer.de обсуждение исходный текст |
Ответ на | Non-pausing table scan on 9.6 replica? (Mark Fletcher <markf@corp.groups.io>) |
Ответы |
Re: Non-pausing table scan on 9.6 replica?
|
Список | pgsql-general |
Am 06.03.19 um 01:26 schrieb Mark Fletcher: > Hi All, > > On a 9.6 streaming replica, we do table scans for stats and other > things. During these scans, the replication is paused (the > 'recovering' postgres process has 'waiting' appended to it). We're not > using transactions with these scans. Is there anything we can do to > prevent the pausing? > > have you set ```max_standby_streaming_delay``? The default is 30 seconds, which means that this will be the maximum time allowed for a replication lag caused by a conflicting query. You can use ``hot_standby_feedback = on``, but the downside will be more bloat on the tables. Regards, Andreas -- 2ndQuadrant - The PostgreSQL Support Company. www.2ndQuadrant.com
В списке pgsql-general по дате отправления: