From 7cc5939b63f60f041948c1179b7d16d5cee41911 Mon Sep 17 00:00:00 2001 From: Zsolt Parragi Date: Mon, 17 Aug 2026 11:52:53 +0000 Subject: [PATCH v2] Report single-page checksum failures in pg_stat_database Base backups reported checksum failures to pg_stat_database only for files with more than one failing page. Commit 6b9e875f728 placed the report inside the block emitting the per-file summary WARNING, which was skipped for a single failure. As a result, a backup failing on files with one corrupted page each left checksum_failures untouched. To fix, emit the per-file summary and the pgstat report for any non-zero failure count. The end-of-backup total WARNING had the same off-by-one and is now also emitted for a single failure. --- src/backend/backup/basebackup.c | 13 ++++++------- src/bin/pg_basebackup/t/010_pg_basebackup.pl | 17 +++++++++++++++++ 2 files changed, 23 insertions(+), 7 deletions(-) diff --git a/src/backend/backup/basebackup.c b/src/backend/backup/basebackup.c index fe5ce23aaba..b95f3577870 100644 --- a/src/backend/backup/basebackup.c +++ b/src/backend/backup/basebackup.c @@ -655,12 +655,11 @@ perform_base_backup(basebackup_options *opt, bbsink *sink, if (total_checksum_failures) { - if (total_checksum_failures > 1) - ereport(WARNING, - (errmsg_plural("%lld total checksum verification failure", - "%lld total checksum verification failures", - total_checksum_failures, - total_checksum_failures))); + ereport(WARNING, + (errmsg_plural("%lld total checksum verification failure", + "%lld total checksum verification failures", + total_checksum_failures, + total_checksum_failures))); ereport(ERROR, (errcode(ERRCODE_DATA_CORRUPTED), @@ -1810,7 +1809,7 @@ sendFile(bbsink *sink, const char *readfilename, const char *tarfilename, CloseTransientFile(fd); - if (checksum_failures > 1) + if (checksum_failures > 0) { ereport(WARNING, (errmsg_plural("file \"%s\" has a total of %d checksum verification failure", diff --git a/src/bin/pg_basebackup/t/010_pg_basebackup.pl b/src/bin/pg_basebackup/t/010_pg_basebackup.pl index cfcfdb8b580..b0d8b505dbd 100644 --- a/src/bin/pg_basebackup/t/010_pg_basebackup.pl +++ b/src/bin/pg_basebackup/t/010_pg_basebackup.pl @@ -913,6 +913,15 @@ $node->command_checks_all( 'pg_basebackup reports checksum mismatch'); rmtree("$tempdir/backup_corrupt"); +# The failure must be counted in pg_stat_database even though the file only +# has a single corrupted page. The walsender reports the failures when it +# exits, so poll for the counter. +ok( $node->poll_query_until( + 'postgres', + 'SELECT checksum_failures = 1 FROM pg_stat_database ' + . "WHERE datname = 'postgres';"), + 'checksum failure reported in pg_stat_database'); + # induce further corruption in 5 more blocks $node->stop; for my $i (1 .. 5) @@ -942,6 +951,14 @@ $node->command_checks_all( 'pg_basebackup correctly report the total number of checksum mismatches'); rmtree("$tempdir/backup_corrupt3"); +# The counter accumulates the failures of all three backups, including the +# single-page failure of the second file in the last one. +ok( $node->poll_query_until( + 'postgres', + 'SELECT checksum_failures = 14 FROM pg_stat_database ' + . "WHERE datname = 'postgres';"), + 'checksum failures accumulated in pg_stat_database'); + # do not verify checksums, should return ok $node->command_ok( [ -- 2.54.0