Fix standby bootstrap when standby_primary_slot_name is set

[mdoulabi1]

Summary

• propagate standby_primary_slot_name to the standby bootstrap helper
• stop pg_receivewal before postgres startup and preserve wal_fast
• use restore_command.sh for local wal_fast replay during standby bootstrap
• keep generic replica bootstrap free from standby slot leakage

Problem

When standby_primary_slot_name is configured, the bootstrap helper may not use the slot correctly, or the handoff from bootstrap WAL collection to postgres recovery may fail. This can lead to:

• replication slot conflicts during startup
• missing WAL errors after basebackup
• broken standby bootstrap on large datasets or low-retention upstreams

Test

• local script/harness validation for basebackup.sh
• local script/harness validation for restore_command.sh
• verified deployed standby bootstrap reaches healthy state
• verified previous errors no longer appear:
  • replication slot "..." is active
  • requested WAL segment ... has already been removed