Commit1b2356a

committed

Fix some comments related to logical repslot advancing

confirmed_flush is part of a replication slot's information, but notconfirmed_lsn.Author: Kyotaro HoriguchiDiscussion:https://postgr.es/m/20191226.175919.17237335658671970.horikyota.ntt@gmail.comBackpatch-through: 11

1 parent9e551a1 commit1b2356aCopy full SHA for 1b2356a

File tree

-4

lines changed

-4

lines changed

Lines changed: 4 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -343,8 +343,8 @@ pg_physical_replication_slot_advance(XLogRecPtr moveto)`
`343`	`343`	`/*`
`344`	`344`	`* Helper function for advancing our logical replication slot forward.`
`345`	`345`	`*`
`346`		`- * The slot's restart_lsn is used as start point for reading records,`
`347`		`- *while confirmed_lsn is used as base point for the decoding context.`
	`346`	`+ * The slot's restart_lsn is used as start point for reading records, while`
	`347`	`+ *confirmed_flush is used as base point for the decoding context.`
`348`	`348`	`*`
`349`	`349`	`* We cannot just do LogicalConfirmReceivedLocation to update confirmed_flush,`
`350`	`350`	`* because we need to digest WAL to advance restart_lsn allowing to recycle`
`@@ -514,8 +514,8 @@ pg_replication_slot_advance(PG_FUNCTION_ARGS)`
`514`	`514`	`/*`
`515`	`515`	`* Check if the slot is not moving backwards. Physical slots rely simply`
`516`	`516`	`* on restart_lsn as a minimum point, while logical slots have confirmed`
`517`		`- * consumption up toconfirmed_lsn, meaning that in both cases data older`
`518`		`- * than that is not available anymore.`
	`517`	`+ * consumption up toconfirmed_flush, meaning that in both cases data`
	`518`	`+ *olderthan that is not available anymore.`
`519`	`519`	`*/`
`520`	`520`	`if (OidIsValid(MyReplicationSlot->data.database))`
`521`	`521`	`minlsn=MyReplicationSlot->data.confirmed_flush;`

Comments

(0)