On 2/24/25 2:05 AM, Michael Paquier wrote:
I think that you have the right idea here, avoiding the duplication
of the errdetail() which feels itchy when looking at the patch.
Done this way in the attached patch.
This should have a note for translators that this field refers to a file
name.
Is there a specific way todo this?
--
Benoit Lobréau
Consultant
http://dalibo.com
From 35dd8db22c997e78ba92f63ffca3022b435d7304 Mon Sep 17 00:00:00 2001
From: benoit <benoit.lobr...@dalibo.com>
Date: Fri, 21 Feb 2025 14:09:56 +0100
Subject: [PATCH] Adjust logging for invalid recovery timeline
The original message doesn't mention where the checkpoint was read from.
It refers to the "Latest checkpoint", the terminology is used for some
records from the control file (specifically, the pg_controldata output).
The backup label uses "Checkpoint location".
---
src/backend/access/transam/xlogrecovery.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/src/backend/access/transam/xlogrecovery.c b/src/backend/access/transam/xlogrecovery.c
index 915cb330295..c5160bf5258 100644
--- a/src/backend/access/transam/xlogrecovery.c
+++ b/src/backend/access/transam/xlogrecovery.c
@@ -849,7 +849,8 @@ InitWalRecovery(ControlFileData *ControlFile, bool *wasShutdown_ptr,
ereport(FATAL,
(errmsg("requested timeline %u is not a child of this server's history",
recoveryTargetTLI),
- errdetail("Latest checkpoint is at %X/%X on timeline %u, but in the history of the requested timeline, the server forked off from that timeline at %X/%X.",
+ errdetail("Latest checkpoint in %s is at %X/%X on timeline %u, but in the history of the requested timeline, the server forked off from that timeline at %X/%X.",
+ haveBackupLabel ? "backup_label" : "pg_control",
LSN_FORMAT_ARGS(CheckPointLoc),
CheckPointTLI,
LSN_FORMAT_ARGS(switchpoint))));
--
2.48.1
