Commitb9de00e

committed

doc: document the need to analyze partitioned tables

Autovacuum does not do it.Reported-by: Justin PryzbyDiscussion:https://postgr.es/m/20210913035409.GA10647@telsasoft.comBackpatch-through: 11

1 parent04f0baa commitb9de00eCopy full SHA for b9de00e

File tree

-4

lines changed

-4

lines changed

Lines changed: 9 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -821,10 +821,15 @@ analyze threshold = analyze base threshold + analyze scale factor * number of tu`
`821`	`821`	`</para>`
`822`	`822`
`823`	`823`	`<para>`
`824`		`- Partitioned tables are not processed by autovacuum. Statistics`
`825`		`- should be collected by running a manual <command>ANALYZE</command> when it is`
`826`		`- first populated, and again whenever the distribution of data in its`
`827`		`- partitions changes significantly.`
	`824`	`+ Partitioned tables do not directly store tuples and consequently`
	`825`	`+ are not processed by autovacuum. (Autovacuum does process table`
	`826`	`+ partitions just like other tables.) Unfortunately, this means that`
	`827`	`+ autovacuum does not run <command>ANALYZE</command> on partitioned`
	`828`	`+ tables, and this can cause suboptimal plans for queries that reference`
	`829`	`+ partitioned table statistics. You can work around this problem by`
	`830`	`+ manually running <command>ANALYZE</command> on partitioned tables`
	`831`	`+ when they are first populated, and again whenever the distribution`
	`832`	`+ of data in their partitions changes significantly.`
`828`	`833`	`</para>`
`829`	`834`
`830`	`835`	`<para>`

Comments

(0)