Factorio Field guide

Factorio desync: collect useful diagnostic evidence

Соберите отчёт о desync Factorio, точные версии и сценарий воспроизведения, сохранив приватность мира.

Factorio desync: collect useful diagnostic evidence

When Factorio creates a desync report, preserve it first and note what was happening in the game. The report contains client and server copies of game state, but it does not name the cause. Do not start by removing mods or restarting the only instance: that can erase a reproducible trigger or change the save.

A desync means the server and at least one client have diverged in their simulation. It is not another name for high ping or network lag: network latency alone does not explain a state mismatch. Possible causes include a game, mod/scenario, or hardware-calculation bug.

Collect Factorio desync evidence. The diagram explains the workflow; it is not a live profile or test result.

Preserve the first useful report#

After a desync, wait for the client to finish creating its report. Keep the original ZIP separate and do not edit it. Record the local time and time zone, build number, server game version, and when the issue appeared: immediately on join, after unpausing, or during play.

The report may contain client and server copies of the map and game state, plus log.txt. It can be large and reveal factory layouts, station names, player actions, and private world configuration. Keep the original private. Before posting it publicly, get consent from the map owner and participants; if a written description is enough, do not upload the archive to a public forum or chat.

Write a short incident record#

The administrator and mod author will need:

  • exact client and server versions, and stable or experimental branch;
  • whether participants have Space Age and whether the save uses its content;
  • the complete mod list with exact versions and dependencies; mark recent changes;
  • the scenario and save version, without publishing the save itself without permission;
  • the event sequence before the desync: player actions, planet, train, build, or timing after join/unpause;
  • whether it repeats on a copy and after which exact action;
  • relevant client and server log lines from the same time, with tokens and personal data removed.

Do not call a random block of recent log lines “the desync log.” The lines around the same time and exact versions matter. Review the full log before sharing it: it can reveal a login name and local file paths. Remove those and any other sensitive details that appear.

Look for a trigger on a copy#

Create a separate copy of the world and try the smallest safe reproduction. If the issue began after a mod update, you can roll back that specific mod on the copy and compare with the previous state, keeping both results. If there are many mods, coordinate with their authors before testing groups or removing them one by one on a test map. Record the exact mod set and versions after every step.

Do not disable mods on the live save: entities, recipes, or data they created may disappear or damage the world. If the issue happens only when a new player joins, pausing the server can change when it appears, but that does not establish or fix the cause.

Factorio forums include cases where a recurring join desync was traced to a specific mod configuration. These are useful examples of the investigation method, not a diagnosis of your factory. The same message or trigger timing does not prove the same bug.

Send evidence to someone who can investigate#

For a modded world, contact the mod or scenario author first. For an unmodded game, use the official forum area for game bugs. Ask what minimum evidence they need, and share reports privately when they contain map data.

Comparing report files and Heavy mode may help a developer narrow the search. But internal dumps are not always readable as ordinary text, and differences between saves are not necessarily the root cause: the server copy may have been captured after further changes. Do not declare a mod responsible based on one diff.

A restart can sometimes let players rejoin temporarily, but the triggering conditions may return; it is not a guaranteed fix. Preserve the report, logs, and exact configuration first. If the symptom disappears after a restart, record that as an observation and continue testing a reproduction on a copy.

Sources#

A report is diagnostic evidence, not an automatic fix. Do not publish it until you have reviewed what it reveals about the factory and players.

If the server runs on your Linux VPS, Nexardo can help manage your own infrastructure — learn about the product.