[lesnik512]

docker-compose.dr.yml starts two standbys, pg-replica and pg-standby, and each runs pg_basebackup -X stream against the primary at the same time. The second backup forces a checkpoint, and with no wal_keep_size the primary is free to recycle a WAL segment the first backup is still streaming. That backup then dies and the standby container exits 1:

outbox-dr-replica  | pg_basebackup: error: unexpected termination of replication stream: ERROR:  requested WAL segment 000000010000000000000002 has already been removed
outbox-dr-replica  | pg_basebackup: error: background process terminated unexpectedly

This PR starts the primary with wal_keep_size=256MB, so it holds enough WAL for both backups to finish. db-retry hit the same race in modern-python/db-retry#84 and fixed it with the same setting.

Each start below ran docker compose -f docker-compose.dr.yml down --remove-orphans --volumes, then up -d --wait pg-replica pg-standby, and grepped the standby logs for "has already been removed".

	Starts	Failed	WAL segment removed
main	15	7	7
this PR	15	0	0

Every failure on main was this error. With the fix, just test-dr passes both scenarios (kill and blackhole), and just lint is clean.

[github-actions]

Benchmark gate

✅ gate passed

scenario	msg/s	delete/msg	WALrec/msg	WALB/msg	fpi	upd	del	dead_tup
consumer/w1/b10	1313	1.000	6.72	922	242	5000	5000	10000
consumer/w1/b100	1434	1.000	7.11	1154	243	5000	5000	10000
consumer/w2/b10	1574	1.000	6.71	922	242	5000	5000	10000
consumer/w2/b100	1926	1.000	7.11	1154	243	5000	5000	10000
consumer/w4/b10	1621	1.000	6.76	950	242	5000	5000	10000
consumer/w4/b100	1957	1.000	7.11	1153	243	5000	5000	10000
consumer/w1/b100/tfbs100	5651	0.010	6.11	1130	243	5000	5000	10000
consumer/w1/b100/interval1-10	1445	1.000	7.17	1196	265	5000	5000	10000
producer/w1/b100	3021	0.000	3.04	597	1	0	0	0
producer/w1/b100/ppt1	1015	0.000	4.04	635	0	0	0	0

Gated (fails the build): delete_calls + tuple counters (upd/del/ins) + the producer's insert_calls, exact; select_calls and other_calls within +2; wal_records within a 10% band; fetch_calls within 2x. msg/s, WAL bytes and total calls are informational (timing/FPI noise).