Re: doc: Improve wal_level and effective_wal_level GUC around logical replication - Mailing list pgsql-hackers

From David G. Johnston
Subject Re: doc: Improve wal_level and effective_wal_level GUC around logical replication
Date
Msg-id CAKFQuwbmDM_X44PwxUw2=och-cs1TMMtBvNPuaB_u5EBbaiG2A@mail.gmail.com
Whole thread Raw
In response to Re: doc: Improve wal_level and effective_wal_level GUC around logical replication  (shveta malik <shveta.malik@gmail.com>)
List pgsql-hackers
Thank you, and the others, for the reviews.

On Tuesday, April 7, 2026, shveta malik <shveta.malik@gmail.com> wrote:
Few comments:

1)
+        Running read-only queries on a standby server.

Looks incomplete and disconnected from previous sentence.

Yeah, I missed that one.
 

2)
+        level used by the system.  Namely, whether replica has been promoted to
+        logical (minimal never promotes).

We shall replace replica, logical, minimal with below
<literal>replica</literal>
<literal>logical</literal>
<literal>minimal</literal>

Yeah, I need to do a markup pass too apparently.  Though I do question whether we are being too rote about these.  In this sentence yes I use the actual labels but it’s more about concept than talking about something you’d literally type somewhere.  One doesn’t promote a literal, they promote the operating mode of the server which is then reflected by the change of a value.
 

Also shall we add 'wal_level'? Otherwise, 'replica' on its own could
be understood as a replica server:

Namely, whether wal_level replica has been promoted to ....

 
The preceding sentence to which “namely” links talks about the levels.  But I’ll give it a second look.  From my reply above, writing “WAL level has been promoted to logical from replica”, is probably a better structure.
 
3)
+        The <xref linkend="guc-effective-wal-level"/> parameter
reports the actual
+        level used by the system.  Namely, whether replica has been promoted to
+        logical (minimal never promotes).
+        This parameter can only be set at server start.

"This" here is misleading. Does this refer to wal_level or effective_wal_level?

When talking about where/when a setting can specified, “this” always refers to the setting being documented.  I’m doubtful of the need to reword things so that some other setting being talked about doesn’t get confused instead.  The fact that we say effective WAL level is basically a runtime derived setting in the prior sentence reinforces this belief.  And we try to keep this sentence consistently at the end of the description.  Maybe it needs its own paragraph though?  I’ll consider that.
 

4)
Below that is <literal>minimal</literal>,

Shall we change "that" to <literal>replica</literal> to avoid any confusion?

I’m on the fence at the moment.  This entire section needs a relook anyway and I’ll keep all these in mind.

David J.

pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: Stack-based tracking of per-node WAL/buffer usage
Next
From: Peter Smith
Date:
Subject: Re: Logical Replication - revisit `is_table_publication` function implementation