Sleep after killing server in replication test

Anders Åstrand · Anders Åstrand · commit 557f355fc791 · 2025-05-19T19:13:12.000+02:00
This test often failed in CI, and I believe the cause is that we tried
to start the cluster again before it had been fully killed.
diff --git a/contrib/pg_tde/t/012_replication.pl b/contrib/pg_tde/t/012_replication.pl
@@ -75,6 +75,7 @@
 PGTDE::psql($primary, 'postgres',
 	"ALTER SYSTEM SET pg_tde.wal_encrypt = 'on';");
 $primary->kill9;
+sleep 1; # Make sure we don't try to start the server again before it's been properly killed.
 
 PGTDE::append_to_result_file("-- primary start");
 $primary->start;
