Limits and gotchas: migrate from Valkey to KiviDB

What does not come across

• Valkey 9 snapshots, by any method that reads them directly. Use RedisShake sync_reader.
• Search indexes and Bloom filters. Run FT.CREATE again; rebuild Bloom filters from the source data.
• Sentinel. There is none in KiviDB.
• Configuration and users, which are not in a snapshot. Set them on KiviDB.

Gotchas

• Key names must be valid UTF-8. KiviDB refuses other key names in SET, RESTORE and the commands copy tools send (ERR key must be valid UTF-8), and RIOT stops at the first one. Check before you copy:

  python3 - "redis://:$PASSWORD@source.internal:6379" <<'EOF'
  import sys, redis
  bad = 0
  for db in range(16):  # the databases your source uses
      r = redis.Redis.from_url(sys.argv[1], db=db)
      for key in r.scan_iter(count=1000):
          try:
              key.decode("utf-8")
          except UnicodeDecodeError:
              bad += 1
              print(f"db{db} {key!r}")
  print(f"{bad} key names are not valid UTF-8")
  EOF
  

  Rename those keys at the source, or leave them out of the copy.

• A KiviDB replica of a Valkey 9 primary keeps retrying the full sync and stays empty. Watch for not a Redis RDB file in KiviDB's log.

• RedisShake stops at the first hash with field expiry (ERR Bad data format). Filter those keys out first.

• Tested with Valkey 8.1.11 (replication, RDB file, --preload-file, MIGRATE, RedisShake sync_reader and scan_reader, RIOT) and Valkey 9.0.7 and 9.1.3 (RedisShake sync_reader with the field-expiry script, RIOT --struct), with every core data type, TTLs, hash field expiry, streams with consumer groups and a second logical database.