Commitf1780ea

committed

doc: pg_upgrade, clarify standby servers must remain running

Also mention that mismatching primary/standby LSNs should neverhappen.Reported-by: Nikolay SamokhvalovDiscussion:https://postgr.es/m/CAM527d8heqkjG5VrvjU3Xjsqxg41ufUyabD9QZccdAxnpbRH-Q@mail.gmail.comBackpatch-through: 11

1 parent28919a0 commitf1780eaCopy full SHA for f1780ea

File tree

-4

lines changed

-4

lines changed

Lines changed: 2 additions & 4 deletions

Original file line number	Diff line number	Diff line change
`@@ -354,8 +354,8 @@ NET STOP postgresql-&majorversion;`
`354`	`354`	`</para>`
`355`	`355`
`356`	`356`	`<para>`
`357`		`- Streaming replication and log-shipping standby serverscan`
`358`		`-remainrunninguntil a later step.`
	`357`	`+ Streaming replication and log-shipping standby serversmust be`
	`358`	`+ runningduring this shutdown so they receive all changes.`
`359`	`359`	`</para>`
`360`	`360`	`</step>`
`361`	`361`
`@@ -368,8 +368,6 @@ NET STOP postgresql-&majorversion;`
`368`	`368`	`servers are caught up by running <application>pg_controldata</application>`
`369`	`369`	`against the old primary and standby clusters. Verify that the`
`370`	`370`	`<quote>Latest checkpoint location</quote> values match in all clusters.`
`371`		`- (There will be a mismatch if old standby servers were shut down`
`372`		`- before the old primary or if the old standby servers are still running.)`
`373`	`371`	`Also, make sure <varname>wal_level</varname> is not set to`
`374`	`372`	`<literal>minimal</literal> in the <filename>postgresql.conf</filename> file on the`
`375`	`373`	`new primary cluster.`

Comments

(0)