Skip to main content

KnotLink and Integration Mod

MineRewind hot backup and hot restore depend on KnotLink and a Minecraft integration mod. A FolderRewind 1.8 environment needs KnotLink Server v3, and every FolderRewind/MineRewind command uses strict key-value parameterized protocol v2.

Prerequisites​

  • FolderRewind 1.8 and MineRewind satisfying its MinHostVersion.
  • A compatible integration mod that handles handshake, save, exit, and rejoin events.
  • A running KnotLink Server v3 with working send and receive paths.

Send cmd=GET_CAPABILITIES first. If the runtime manifest does not include MineRewind's current_save capabilities, do not continue with current-world commands.

Current-world commands​

MineRewind extends the Host's BACKUP, LIST_BACKUPS, and RESTORE through parameters. It no longer uses a second set of space-delimited commands:

Purposev2 request
Back up current worldcmd=BACKUP;current_save=true;from=minebackup.mod;request_id=...
List current-world backupscmd=LIST_BACKUPS;current_save=true
Restore latest backupcmd=RESTORE;current_save=true;from=minebackup.mod;request_id=...
Restore a selected backupAdd file=<encoded filename> to the previous request
Preserve player dataAlso add preserve_player_data=true

Current-world backup can also use the Host's one-shot comment, backup_mode, compression_method, and compression_level overrides. They affect only this invocation.

cmd=BACKUP;current_save=true;comment=QuickSave;backup_mode=full;from=minebackup.mod;request_id=mc-backup-001

Integration callbacks​

The mod reports stages back to MineRewind using the same v2 format:

cmd=HANDSHAKE_RESPONSE;mod_version=1.8.0
cmd=WORLD_SAVED
cmd=WORLD_SAVE_AND_EXIT_COMPLETE
cmd=REJOIN_RESULT;result=success

Dynamic text such as a failure reason must be percent-encoded:

cmd=REJOIN_RESULT;result=failure;reason=Server%20not%20ready

MineRewind signals​

  • handshake / handshake_ack negotiate versions.
  • pre_hot_backup asks the mod to save before hot backup.
  • hot_restore_requested asks the mod to begin its current-world restore countdown.
  • pre_hot_restore asks it to save and exit the world.
  • restore_cancelled reports a cancelled restore flow.
  • rejoin_world asks it to rejoin the restored world.
  • hot_restore_complete reports the final end-to-end status.

The request_id in hot_restore_requested is reused by the later RESTORE conversation. Correlate one restore by this field, not by event ordering alone.

Minimal integration test​

  1. Send cmd=PING to check the FolderRewind endpoint.
  2. Send cmd=GET_CAPABILITIES to inspect MineRewind capabilities and fields.
  3. Send current-world BACKUP and observe save and backup-completion signals.
  4. Send current-world LIST_BACKUPS and verify the query result.
  5. Run latest-backup RESTORE against a test world and observe save, exit, restore, and rejoin.
  6. Finally test precise restore with file and optional player-data preservation.

:::danger Use a test world first Hot restore changes a world that is in use. A failed version handshake, stage timeout, or continued world writes can cancel the flow. Rehearse it end to end and keep an independently restorable full backup before production use. :::