Before you start
Record the server version, server.identity, server.seed, server.worldsize and installed plugins. Reserve a maintenance window and storage outside the working directory. File access and backups do not imply that an automatic wipe scheduler is included.
Define the reset
A map wipe replaces the world and its buildings. A blueprint wipe resets learned crafting recipes. Plugins may hold separate economy, permission or progression data. Write a clear retention list before maintenance: new world, retained or cleared blueprints, and retained or migrated plugin data. Your community schedule must also account for update requirements announced by Facepunch.
Changing hosts is not a wipe decision. Stop writes and preserve the actual server.identity directory with world saves, player data and blueprints from the same coherent backup. Keep configuration, uMod/Carbon plugins and their data, and any external SQL separately. Restore the same compatible server/plugin versions privately, test progression after a restart and agree any deliberate world or blueprint wipe before switching.
- Record exactly what resets: the map, blueprints, plugin economy, permissions and other progression.
- Tell players the time zone, estimated interruption and expected data loss; allow them to disconnect before maintenance.
Stop writes and preserve a coherent backup
Capture the backup after a clean stop and confirm in the logs that the process exited. Rust SQLite databases can have -wal files: copying only a .db while writes continue is not a coherent backup. Keep the identity directory, configuration and framework data together.
Open the archive in a separate location to check its contents. A backup name or a non-zero size does not prove recoverability. Record the file count, capture time, server and plugin versions, and keep the recovery copy outside the working wipe directory.
Identify world and blueprint files
Inside server/<identity>, .map and .sav files belong to the world; blueprint data is separate. Exact filenames depend on the current release. Inspect the real directory and official documentation instead of deleting every .db file. Move selected files into a recovery copy rather than destroying them immediately. Do not rename an old save merely to force it through an incompatible version change.
Read-only Linux file inventory
find './server/my-identity' -maxdepth 1 -type f -printf '%f\n'Run from the actual Rust directory and replace my-identity. This lists names without changing files. Identify this version’s .sav, .map and player.blueprints files; the command does not decide which files should be removed.
Prepare the map and plugin environment
For a procedural map, record the intended seed and worldsize together. Larger maps, structures and entities can increase memory, CPU and disk usage. Vanilla needs no mod framework. uMod/Oxide and Carbon have their own plugin compatibility and update requirements. Documentation links are not a promise of a panel installer. Validate framework changes separately from the wipe whenever possible.
Verify the first startup
Start during the maintenance window and inspect the first meaningful log errors. Check the actual map, save handling and administrator permissions. Connect with a suitable test account and verify blueprint retention. Test problem plugins separately on a copy. A working RCON connection proves access to an administration interface; it does not establish that the world or every gameplay extension is healthy.
- Confirm the new map loads and the logs have no repeating plugin failures.
- Connect with a test account and check the seed, size, spawn, permissions and persistence after a save.
- Separately test a returning player whose blueprints should remain and a fresh player; check framework economy if used.
Reopen with a recovery plan
Agree rollback criteria before reopening: the wrong map, inconsistent progression or failures that prevent play. Stop writes again and restore the coherent backup set rather than a world file copied from another point in time. Recovery can lose actions taken after reopening; explain this limit to the community.
Keep the original files until validation and record the parameters actually applied. Do not bypass a protocol-forced wipe by renaming an old save. Plugin data and the Rust+ identity are not disposable merely because they share a directory with world files.
For your situation
| Situation | Next check |
|---|---|
| Keep blueprints | Reset only the identified world data after backup, keeping the separate blueprint database. |
| Switch frameworks | Test plugin compatibility and migration separately instead of stacking unverified changes. |
| World fails to load | Read the first relevant error, check versions and return to a coherent recovery copy. |
Official sources
These documents describe the relevant protocols and software. Your service terms and configuration still need to be checked.