Re: v13: CLUSTER segv with wal_level=minimal and parallel index creation

This is easily reproduced, at least on pg_attribute

[pryzbyj(at)localhost ~]$ /usr/pgsql-13/bin/initdb -D pgsql13.dat
[pryzbyj(at)localhost ~]$ /usr/pgsql-13/bin/postgres -D pgsql13.dat -c logging_collector=off -c port=5678 -c unix_socket_directories=/tmp -c wal-level=minimal -c max_wal_senders=0&
[pryzbyj(at)localhost ~]$ psql -h /tmp -p 5678 postgres -c "SET min_parallel_table_scan_size=0" -c "SET client_min_messages=debug" -c "REINDEX INDEX pg_attribute_relid_attnum_index"
SET
SET
DEBUG: building index "pg_attribute_relid_attnum_index" on table "pg_attribute" with request for 1 parallel worker
TRAP: FailedAssertion("relcache_verdict == RelFileNodeSkippingWAL(relation->rd_node)", File: "relcache.c", Line: 2976)
postgres: parallel worker for PID 9637 (ExceptionalCondition+0x66)[0x921d86]
postgres: parallel worker for PID 9637 (AssertPendingSyncs_RelationCache+0x1db)[0x91a08b]
postgres: parallel worker for PID 9637 (smgrDoPendingSyncs+0x71)[0x58e591]
postgres: parallel worker for PID 9637 [0x53b1a4]
postgres: parallel worker for PID 9637 (EndParallelWorkerTransaction+0x16)[0x53b826]
postgres: parallel worker for PID 9637 (ParallelWorkerMain+0x437)[0x52fcf7]
postgres: parallel worker for PID 9637 (StartBackgroundWorker+0x25a)[0x76047a]
postgres: parallel worker for PID 9637 [0x76d88d]
postgres: parallel worker for PID 9637 [0x76e43e]
/lib64/libpthread.so.0(+0xf630)[0x7f46d4b47630]
/lib64/libc.so.6(__select+0x13)[0x7f46d26ba983]
postgres: parallel worker for PID 9637 [0x4887bc]
postgres: parallel worker for PID 9637 (PostmasterMain+0x1165)[0x76fb45]
postgres: parallel worker for PID 9637 (main+0x70d)[0x48a83d]
/lib64/libc.so.6(__libc_start_main+0xf5)[0x7f46d25e7555]

--
Justin