Commit1c15aac

committed

Add comment & docs about no vacuum truncation with sto.

Omission noted by Andres Freund.

1 parentb11e9bb commit1c15aacCopy full SHA for 1c15aac

File tree

+18

-0

lines changed

+18

-0

lines changed

Lines changed: 9 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -2091,6 +2091,15 @@ include_dir 'conf.d'`
`2091`	`2091`	`transaction ID wraparound may occur in much shorter time frames.`
`2092`	`2092`	`</para>`
`2093`	`2093`
	`2094`	`+ <para>`
	`2095`	`+ When this feature is enabled, freed space at the end of a relation`
	`2096`	`+ cannot be released to the operating system, since that could remove`
	`2097`	`+ information needed to detect the <literal>snapshot too old</>`
	`2098`	`+ condition. All space allocated to a relation remains associated with`
	`2099`	`+ that relation for reuse only within that relation unless explicitly`
	`2100`	`+ freed (for example, with <command>VACUUM FULL</>).`
	`2101`	`+ </para>`
	`2102`	`+`
`2094`	`2103`	`<para>`
`2095`	`2104`	`This setting does not attempt to guarantee that an error will be`
`2096`	`2105`	`generated under any particular circumstances. In fact, if the`

Lines changed: 9 additions & 0 deletions

Original file line number	Diff line number	Diff line change
`@@ -1663,6 +1663,15 @@ lazy_cleanup_index(Relation indrel,`
`1663`	`1663`	`* Don't even think about it unless we have a shot at releasing a goodly`
`1664`	`1664`	`* number of pages. Otherwise, the time taken isn't worth it.`
`1665`	`1665`	`*`
	`1666`	`+ * Also don't attempt it if we are doing early pruning/vacuuming, because a`
	`1667`	`+ * scan which cannot find a truncated heap page cannot determine that the`
	`1668`	`+ * snapshot is too old to read that page. We might be able to get away with`
	`1669`	`+ * truncating all except one of the pages, setting its LSN to (at least) the`
	`1670`	`+ * maximum of the truncated range if we also treated an index leaf tuple`
	`1671`	`+ * pointing to a missing heap page as something to trigger the "snapshot too`
	`1672`	`+ * old" error, but that seems fragile and seems like it deserves its own patch`
	`1673`	`+ * if we consider it.`
	`1674`	`+ *`
`1666`	`1675`	`* This is split out so that we can test whether truncation is going to be`
`1667`	`1676`	`* called for before we actually do it. If you change the logic here, be`
`1668`	`1677`	`* careful to depend only on fields that lazy_scan_heap updates on-the-fly.`

Comments

(0)