From a309bff820b9c3a1cf178388eef4bf229888cb1e Mon Sep 17 00:00:00 2001 From: Kyotaro Horiguchi Date: Wed, 18 Aug 2021 13:58:15 +0900 Subject: [PATCH v2] Fix basebackup to generate correct WAL-Ranges info. While WAL-Ranges in backup_manifest is supposed to have the WAL range to apply at recovery. it starts from the beginning of the oldest timeline if the starting TLI > 1. As the result pg_verifybackup may result in a false failure for a valid backup. Adjust the logic to make the WAL-Ranges starts with the actual start LSN. Backpatch to PG13 where this feature has been introduced. --- src/backend/replication/backup_manifest.c | 17 ++++++++++++----- src/bin/pg_verifybackup/t/007_wal.pl | 18 +++++++++++++++++- 2 files changed, 29 insertions(+), 6 deletions(-) diff --git a/src/backend/replication/backup_manifest.c b/src/backend/replication/backup_manifest.c index 1e07d99e64..a672690c7d 100644 --- a/src/backend/replication/backup_manifest.c +++ b/src/backend/replication/backup_manifest.c @@ -236,11 +236,18 @@ AddWALInfoToBackupManifest(backup_manifest_info *manifest, XLogRecPtr startptr, errmsg("expected end timeline %u but found timeline %u", starttli, entry->tli)); - if (!XLogRecPtrIsInvalid(entry->begin)) - tl_beginptr = entry->begin; - else - { + /* + * If this timeline entry matches with the timeline on which the backup + * started, WAL needs to be checked from the start LSN of the backup. + * If this entry refers to a newer timeline, WAL needs to be checked + * since the beginning of this timeline, so use the LSN where the + * timeline began. + */ + if (starttli == entry->tli) tl_beginptr = startptr; + else + { + tl_beginptr = entry->begin; /* * If we reach a TLI that has no valid beginning LSN, there can't @@ -248,7 +255,7 @@ AddWALInfoToBackupManifest(backup_manifest_info *manifest, XLogRecPtr startptr, * better have arrived at the expected starting TLI. If not, * something's gone horribly wrong. */ - if (starttli != entry->tli) + if (!XLogRecPtrIsInvalid(entry->begin)) ereport(ERROR, errmsg("expected start timeline %u but found timeline %u", starttli, entry->tli)); diff --git a/src/bin/pg_verifybackup/t/007_wal.pl b/src/bin/pg_verifybackup/t/007_wal.pl index 56d536675c..ae35e035eb 100644 --- a/src/bin/pg_verifybackup/t/007_wal.pl +++ b/src/bin/pg_verifybackup/t/007_wal.pl @@ -7,7 +7,7 @@ use Config; use File::Path qw(rmtree); use PostgresNode; use TestLib; -use Test::More tests => 7; +use Test::More tests => 9; # Start up the server and take a backup. my $master = get_new_node('master'); @@ -56,3 +56,19 @@ command_fails_like( [ 'pg_verifybackup', $backup_path ], qr/WAL parsing failed for timeline 1/, 'corrupt WAL file causes failure'); + +# +# Check that WAL-Ranges has the correct values when TLI > 1 +$master->stop; +$master->append_conf('standby.signal'); +$master->start; +$master->promote; +$master->safe_psql('postgres', 'SELECT pg_switch_wal()'); +my $backup_path2 = $master->backup_dir . '/test_tli'; +# base-backup below runs a checkpoint, which removes the first segment +# of the current timeline +$master->command_ok([ 'pg_basebackup', '-D', $backup_path2, '--no-sync' ], + "base backup 2 ok"); +# Should pass this case +command_ok([ 'pg_verifybackup', $backup_path2 ], + 'verifybackup should pass for a valid backup on timeline > 1'); -- 2.27.0