Skip to content

Server owners

Configure work budgets, understand persistence and support players with rules.

  1. Keep the old mod version installed while every owner finishes or cancels unfinished paid jobs. A paused job is still unfinished.
  2. Make eligible return space available and settle all owed unused-input returns on that old version. Resolve any recovery condition before treating the work as closed.
  3. Stop the server or singleplayer instance cleanly and back up the complete world, including its schematic job data and server configuration.
  4. Install the new JAR matching the Minecraft version and loader on both server and clients, then reopen the world.
  5. Check Minecraft Controls: an old 0.1.0 K menu binding can remain after updating; 0.2.0 uses V by default.

Saved paid work records the runtime environment, including the mod version. Unfinished work crossing a version change can enter Recovery with an environment-changed message. An update does not discard its records, complete the job or settle an owed return.

Downgrade by restoring the pre-upgrade world backup and its matching old mod files. Removing 0.2.0 and opening the same upgraded world with 0.1.0 is unsupported: printer blocks and saved printer templates exist independently of active jobs and cannot be loaded by 0.1.0.

Players can follow the getting-started upgrade checklist. Printer owners should also review Stop, returns and restart behavior.

Install the same mod build on server and clients. The world’s serverconfig/ folder contains:

Loader Configuration
NeoForge / Forge simpleschematics-server.toml
Fabric simpleschematics-server.properties

Defaults allow 4 ms / 8,192 Creative steps per tick, 2 ms of shared Survival work per tick and up to 200 paid groups per game second. Undo budgets default to 64 MiB per player and 512 MiB per world, measured as decoded data. Actual work is also limited by journal and storage operations.

Players can view rules; operator permission level 2 is required for world-rule edits. Datapack rules reload with /reload. The Rules guide covers field discovery, costs, precedence and invalid files.

Check claim compatibility on your server. Native events and Simple Schematics’ extension events support permission decisions; private claim APIs may require integration.

There is no fixed 4,096-block or one-million-block cap. Whether a schematic fits depends on its format, palette, saved data, occupied sections and available memory. The budgets below apply independently of block count.

Budget Limit
Compressed input 64 MiB
Decoded NBT 512 MiB
Section arrays 256 MiB
Estimated placement working storage 256 MiB
Cached preview GPU geometry 128 MiB
One saved block / complete entity group 256 KiB
Root entities 4,096
Passenger group 256 members; 16 nesting edges

The root-entity limit is not a block limit. Sparse regions do not allocate every empty cell in their bounding volume. Fit depends on source format, palette, saved data and memory.

World mutations run on the server thread. Parsing and history I/O use workers. Versioned journals are written before mutations. Interrupted or ambiguous operations keep evidence for the recovery inspector and require fresh validation before resolution.

A clean reload that offers the printer owner Resume is different from an ambiguous operation after a crash or interrupted write. If the screen reports Recovery, inspect the recorded state first and follow the available recovery actions. Do not assume a second start or a refund is safe solely because the server is back online.

Unknown schemas can be archived explicitly while preserving their original bytes. Old jobs with verified ownership can remain closable in Recovery after leaving the recent-history index. Closing preserves evidence and does not refund items.