Fix more race conditions in the newly-added pg_rewind test.

pg_rewind looks at the control file to check what timeline a server is on.
But promotion doesn't immediately write a checkpoint, it merely writes
an end-of-recovery WAL record. If pg_rewind runs immediately after
promotion, before the checkpoint has completed, it will think think that
the server is still on the earlier timeline. We ran into this issue a long
time ago already, see commit 484a848a73f.

It's a bit bogus that pg_rewind doesn't determine the timeline correctly
until the end-of-recovery checkpoint has completed. We probably should
fix that. But for now work around it by waiting for the checkpoint
to complete before running pg_rewind, like we did in commit 484a848a73f.

In the passing, tidy up the new test a little bit. Rerder the INSERTs so
that the comments make more sense, remove a spurious CHECKPOINT call after
pg_rewind has already run, and add --debug option, so that if this fails
again, we'll have more data.

Per buildfarm failure at https://buildfarm.postgresql.org/cgi-bin/show_stage_log.pl?nm=rorqual&dt=2020-12-06%2018%3A32%3A19&stg=pg_rewind-check.
Backpatch to all supported versions.

Discussion: https://www.postgresql.org/message-id/1713707e-e318-761c-d287-5b6a4aa807e8@iki.fi

1 files changed, 15 insertions, 7 deletions

diff --git a/src/bin/pg_rewind/t/008_min_recovery_point.pl b/src/bin/pg_rewind/t/008_min_recovery_point.pl
index 4968acadc4f..7330833ac00 100644
--- a/src/bin/pg_rewind/t/008_min_recovery_point.pl
+++ b/src/bin/pg_rewind/t/008_min_recovery_point.pl
@@ -75,6 +75,13 @@ $node_1->wait_for_catchup('node_3', 'replay', $lsn);
 #
 $node_1->stop('fast');
 $node_3->promote;
+# Force a checkpoint after the promotion. pg_rewind looks at the control
+# file to determine what timeline the server is on, and that isn't updated
+# immediately at promotion, but only at the next checkpoint. When running
+# pg_rewind in remote mode, it's possible that we complete the test steps
+# after promotion so quickly that when pg_rewind runs, the standby has not
+# performed a checkpoint after promotion yet.
+$node_3->safe_psql('postgres', "checkpoint");
 
 # reconfigure node_1 as a standby following node_3
 my $node_3_connstr = $node_3->connstr;
@@ -105,6 +112,8 @@ $lsn = $node_3->lsn('insert');
 $node_3->wait_for_catchup('node_1', 'replay', $lsn);
 
 $node_1->promote;
+# Force a checkpoint after promotion, like earlier.
+$node_1->safe_psql('postgres', "checkpoint");
 
 #
 # We now have a split-brain with two primaries. Insert a row on both to
@@ -112,6 +121,9 @@ $node_1->promote;
 # see the insert on 1, as the insert on node 3 is rewound away.
 #
 $node_1->safe_psql('postgres', "INSERT INTO public.foo (t) VALUES ('keep this')");
+# 'bar' is unmodified in node 1, so it won't be overwritten by replaying the
+# WAL from node 1.
+$node_3->safe_psql('postgres', "INSERT INTO public.bar (t) VALUES ('rewind this')");
 
 # Insert more rows in node 1, to bump up the XID counter. Otherwise, if
 # rewind doesn't correctly rewind the changes made on the other node,
@@ -120,10 +132,6 @@ $node_1->safe_psql('postgres', "INSERT INTO public.foo (t) VALUES ('keep this')"
 $node_1->safe_psql('postgres', "INSERT INTO public.foo (t) VALUES ('and this')");
 $node_1->safe_psql('postgres', "INSERT INTO public.foo (t) VALUES ('and this too')");
 
-# Also insert a row in 'bar' on node 3. It is unmodified in node 1, so it won't get
-# overwritten by replaying the WAL from node 1.
-$node_3->safe_psql('postgres', "INSERT INTO public.bar (t) VALUES ('rewind this')");
-
 # Wait for node 2 to catch up
 $node_2->poll_query_until('postgres',
 	q|SELECT COUNT(*) > 1 FROM public.bar|, 't');
@@ -145,9 +153,10 @@ command_ok(
     [
         'pg_rewind',
         "--source-server=$node_1_connstr",
-        "--target-pgdata=$node_2_pgdata"
+	    "--target-pgdata=$node_2_pgdata",
+	    "--debug"
     ],
-	'pg_rewind detects rewind needed');
+	'run pg_rewind');
 
 # Now move back postgresql.conf with old settings
 move(
@@ -159,7 +168,6 @@ $node_2->start;
 # Check contents of the test tables after rewind. The rows inserted in node 3
 # before rewind should've been overwritten with the data from node 1.
 my $result;
-$result = $node_2->safe_psql('postgres', 'checkpoint');
 $result = $node_2->safe_psql('postgres', 'SELECT * FROM public.foo');
 is($result, qq(keep this
 and this