Palworld

Restore a Palworld world without mixing save data

Keep the current files, restore a complete save set to a test copy, and check characters, guilds and bases before replacing your live Palworld world.

Lyleen standing among flowers in a sunlit Palworld forest
Palworld press image © Pocketpair.

You restore a Palworld backup, connect, and face a new-character screen. Stop there. Creating a replacement character and trying another archive can add fresh changes to an already confusing recovery. First establish which world the server loaded and which save set you restored.

Preserve the files you have now

Close access while you investigate and shut the server down. Keep a separate copy of its current save data, configuration and logs, including any files created during the failed restore. Name it as the current state, not as a known-good backup. You may need it to recover progress that a previous archive lacks.

Record when players last saw the correct world, when the problem began, and the time of each candidate backup. Use one time zone for that record. An archive named with local time and a log recorded in UTC can appear to contradict each other by an hour.

For future backups, Pocketpair documents the /Save command and a timed /Shutdown in its command reference. A saved, stopped copy gives you a clearer recovery point than a file transfer taken during active play. Do not force-stop a healthy server just to save a minute.

Choose one complete recovery point

Use the world and player data from the same backup set. Avoid combining an older world with a newer collection of player files because those files happen to be easier to find. Your recovery needs a consistent point in time, not the most recent timestamp from each folder.

Inspect the archive before extracting it. Check that it contains the expected data rather than only a configuration file, an empty directory, or another archive. Confirm the transfer completed and that you have enough free space for both the preserved current copy and the restored copy.

Pocketpair's configuration reference documents bIsUseBackupSaveData, which creates backups within the save-data directory. Those local backups can provide another recovery candidate. They still share the server's storage, so retain a separate off-machine backup for storage failures or accidental removal.

Restore to an isolated copy first

Prepare a separate test instance or ask support for a safe recovery workspace. Use separate storage and ports. Keep ordinary players out, and make sure no scheduled process can copy its files back over production. Do not point two running instances at the same writable save directory.

Match the original game's server version and mod setup where available, and record any difference you cannot reproduce. Restoring an old save into a different build may introduce a second problem. Keep the original archive untouched so you can attempt a different recovery without relying on the modified test files.

Check the active launch settings and the loaded world rather than assuming the archive's folder name selects it. Pocketpair distinguishes the listening port from public advertisement settings in its server arguments guide; use a different listening port for a test server. If you cannot identify the active save location, collect the logs and ask for help before replacing more files.

Check the players as well as the landscape

Ask a small test group to use their normal accounts. Visit a known base and check ownership, guild membership, storage contents and a sample of Pals. Each tester should confirm their character, equipment and progress. Write down what matches and what differs.

A restored landscape with missing characters needs further investigation. Do not assume a crossplay connection setting converts account identities or save formats. Pocketpair currently lists allowed platforms under CrossplayPlatforms; allowing a platform to join does not establish that a save transfer between editions will preserve its characters. Keep Steam and Microsoft/Xbox migration questions separate from an ordinary restore on the same server.

Save the test instance, shut it down and start it again. Reconnect before approving the recovery. That second start can reveal a wrong active directory or a permission problem that the first session did not expose.

Once the test passes, arrange a final maintenance window and preserve production again before applying the approved recovery. Tell players the recovery timestamp and the progress they should expect to lose. Our Palworld world-management article helps document those rules; our backup-retention guide covers the space needed to keep recovery choices.

Official references