Stop the restart loop before changing settings: preserve the save and logs, identify whether the process is crashing or hanging, then test memory, files, mods, saves, and service permissions in that order. This method isolates the cause without sacrificing the only working world.
Godlike.Host offers hosted ARK: Survival Evolved servers with its own game panel, SFTP access and free DDoS protection, if you would rather not host it yourself.
Use this order when the server is crashing
ARK failures that look identical to players can have very different causes. A process that exits, an engine that remains open but stops responding, a mod download failure, a map-load crash, and an out-of-memory termination require different remedies. Work through the following sequence and keep each change reversible.
As an Amazon Associate I earn from qualifying purchases.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →- Protect the data. Stop the server once, then copy the complete
ShooterGame/Saveddirectory to a separate location. Include configuration, cluster data, and every existing crash dump or log. Do not delete or overwrite the original while testing. - Capture evidence. Open
ShooterGame/Saved/Logs/ShooterGame.logand the newest crash dump before starting the service repeatedly. Repeated launches can replace the runtime output that explains the first failure. - Classify the symptom. Record whether the process exits, stays present with no connections, fails while downloading or loading a mod, fails while opening a particular map or save, or causes the host to report exhausted memory.
- Validate the installation. After an update or suspected corruption, run SteamCMD’s ARK dedicated-server
app_updateoperation with thevalidateoption. Keep the backup made in step one so you can roll back files or saves. - Run a clean test. Start the same server build with all mods disabled. If it stays online, add the required mods back in small groups rather than restoring the whole list at once.
- Check the host. Watch RAM, CPU, disk space, and swap or pagefile usage while reproducing the fault. Also verify that the service account can read and write the installation and its
ShooterGame/Savedsubdirectory.
Identify the failure before choosing a fix
| What you observe | Most useful evidence | First direction to investigate |
|---|---|---|
| The server process disappears | Crash dump, final lines of ShooterGame.log, operating-system event log |
Memory pressure, corrupt files, incompatible mod, or damaged save |
| The process remains but players cannot connect and CPU falls to zero | Log timestamps, service status, CPU graph, recent configuration or mod changes | Engine hang; preserve evidence before using an automatic restart |
| Failure occurs during mod download or immediately after a mod update | Mod IDs, download messages, server-side mod directory contents | Missing files, mismatched versions, or an incompatible mod |
| Failure occurs while loading one map or existing world | Save name, map name, last successful backup, clean-world test | Corrupt or incompatible save, map state, or persistent mod data |
| Host or operating system reports an out-of-memory event | Peak RAM, swap/pagefile activity, process limits, crash time | Reduce load temporarily and add compatible physical memory or capacity |
An engine hang is not the same as a crash. A watchdog that simply kills and restarts a hung process may restore availability, but it can also hide the trigger and increase the chance of a save rollback. Treat automatic restarts as containment after logs and a backup exist, not as diagnosis.
Protect saves and read the logs first
Make a complete, usable backup
Copy ShooterGame/Saved while the service is stopped. Preserve the directory structure and include the map saves, player and tribe data, configuration files, cluster files, logs, and any crash dumps that are present. Store at least one copy outside the server host. If the server runs several maps, keep the whole Saved tree rather than copying only the currently selected map.
#1 Best Overall
- Capacity – 32GB RAM KIT (2 x 16GB Modules) Speed up to 2400MHz Non-ECC Unbuffered 288-Pin 1.2V UDIMM.
- Specs – Black PCB Color and Dual Rank (2Rx8).
- Compatibility – Designed for selected DDR4 Desktop PCs and workstations that support 288-Pin UDIMM memory. NOT compatible with Laptop SODIMM slots.
- Installation – Plug-and-Play Upgrade, Quick and Easy to Install, no expertise required (please refer to your system's manual for guidelines).
- Warranty – All Timetec products are high-quality and rigorously tested to meet stringent standards. Backed by Timetec Limited Lifetime Warranty and professional technical support based in the United States.
Before changing a build, removing a mod, or restoring a world, create another dated copy. A mod can add persistent items, structures, or classes to a save; removing it before a backup can make recovery harder. Never test by overwriting the only known-good save.
Read the relevant evidence
Start with ShooterGame/Saved/Logs/ShooterGame.log and compare its final timestamp with the operating-system event log. Look for the last completed action: a named mod, a map or save filename, an asset package, an allocation failure, or a permissions error. A crash dump can identify an engine module, but the last successful operation in the runtime log often tells you which change to reverse.
Keep a short incident note containing the server build, map, mod list, player count, time of failure, and what changed immediately beforehand. This prevents a restart loop from erasing the sequence needed to reproduce the problem.
Fix memory pressure and decide whether RAM is the issue
There is no universal minimum or recommended RAM number for every ARK dedicated server. Memory use changes with map choice, connected players, world age, structures, creatures, storage, server configuration, and installed mods. A small new world can run successfully on a host that cannot load an older, heavily built world with the same settings.
Confirm pressure instead of guessing
- Start the server with monitoring active and record total, available, and swap or pagefile memory.
- Watch the process during startup, map loading, player joins, and the period when crashes normally occur.
- Compare the peak with the host’s physical capacity and any container, virtual-machine, or service limit.
- Match the timestamp of an allocation error or operating-system out-of-memory event to the memory graph.
If the graph remains well below the limit and the log names a mod, asset, or save, buying RAM is unlikely to solve the underlying fault. If memory reaches the limit, reduce the immediate workload by stopping extra maps or unnecessary services, then plan a capacity increase.
Choose a compatible server RAM upgrade
When measured memory pressure is confirmed, the upgrade must match the host rather than simply provide a larger number. Check the motherboard or virtual host’s supported RAM generation, maximum capacity, module type, speed, channel layout, and whether ECC is required. Mixing unsupported generations or exceeding the platform limit can prevent booting or create new instability. Recheck memory use after installation with the same map, player load, and mod list.
Free tools Windows power users keep installed
One-click scans. No signup required.
Use automatic restart protection carefully
The server configuration reference documents -GBUsageToForceRestart=<value> as an out-of-memory protection option. It saves the world and restarts when the configured gigabyte threshold is reached. Treat this as a containment measure: it limits the damage from exhaustion but does not increase available memory or fix a leak, oversized world, or incompatible mod. Set it only after confirming that the save operation and service restart work reliably, and keep external logs so repeated threshold restarts are visible.
Rank #2
- Capacity: 16GB (2x 8GB Modules) | Type: DDR3 240-Pin | Speed: 1600MHz PC3-12800 / (PC3-12800E) | ECC Type: ECC-UDIMM (ECC Unbuffered DIMM) | Rank: 2Rx8 (Dual Rank x8) | Voltage: 1.35V
- Designed for ECC UDIMM Compatible Servers/Workstations (Rated Speeds & ECC Capabilities are CPU Dependent). Not Compatible with Desktops/Laptops.
- ECC Types can not be mixed | All installed modules must be ECC UDIMMs in order to function properly | A maximum of eight ranks per memory channel can be installed at once
- All A-Tech memory modules undergo stringent quality control testing to ensure dependable and reliable performance
- Backed by A-Tech's Limited Lifetime Warranty + Tech Support Team available to help before and after your purchase
Isolate mod failures without losing persistent content
A crash immediately after a mod update, while downloading mod content, or during mod initialization strongly suggests a missing file, mismatched version, or mod incompatibility with the current server build. A mod that is present in the configuration but absent from the dedicated server can fail before players ever connect.
Run a controlled no-mod test
- Stop the service and back up the complete save and configuration.
- Record the current mod IDs and their order. Do not rely on memory.
- Temporarily disable every mod in the server’s launch configuration.
- Confirm that the unmodded server can start, load the same map, and remain online long enough to reproduce the normal workload.
- If it is stable, restore a small batch of required mods and repeat. Use a binary split for a long list: add half, test, then split the failing half again.
- When one group fails, verify each ID, confirm that its files actually exist on the server, and update or replace the offending mod only after another backup.
Keep the mod list as small as the server design permits. Every additional mod adds download, startup, memory, and save-compatibility variables. Clients must use versions compatible with the server build; a client-only mismatch can look like a server problem, so compare client connection errors with the dedicated-server log.
Removing a mod that changed the world
If a mod supplied structures, items, creatures, or classes that are already stored in the save, removing it can leave references the world cannot resolve. Preserve the original save, test the removal on a copy, and restore the mod if the copy fails. A clean-world result does not prove the old save is healthy; it only shows that the server executable and current mod set can initialize without that historical data.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Repair a map or save that crashes during loading
When the process fails on one map or one world but other maps start, suspect the save, map state, or persistent mod data before replacing the entire installation. The safest test is a fresh world on the same server build, map, and mod list.
- Copy the affected save and label the copy with the date and build.
- Validate the installation before testing, so a damaged executable is not mistaken for a damaged world.
- Start a fresh world using the same map and, if practical, the same mod configuration.
- If the fresh world starts normally, compare the affected save with the latest known-good backup and investigate the last world, mod, or configuration change.
- Restore the newest backup to a separate test directory and confirm whether the crash follows the save.
Do not overwrite the only copy with a partially loaded or newly generated world. If the backup also fails, test an older backup and then a clean world; this sequence distinguishes progressive save damage from a build-wide problem.
Validate files and recover from a bad update
File corruption is most likely after an interrupted update, failed disk write, or incomplete download. Use SteamCMD’s dedicated-server update workflow with validation before downloading unrelated replacements. The command’s validate option compares installed files and reacquires files that do not match the expected build.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsBack up saves, configuration, and cluster data before validation or any build change. Validation normally targets installation files, but a rollback plan is still essential because a newly updated executable may no longer match an older mod or save. Confirm that the server and connecting clients are on compatible versions after the operation. If validation reports no change and the clean no-mod test still crashes, focus on memory, permissions, or the save rather than repeatedly validating.
Rank #3
- A-Tech RAM Memory compatible for select DDR4 Server and Workstation systems only; (*WILL NOT WORK with Desktop or Laptop Computers/PCs*)
- 32GB RAM Kit (2 x 16GB Modules); DDR4 DIMM 288 Pin; Speeds up to 2400MHz PC4-19200 (PC4-2400T)
- ECC Unbuffered UDIMM; 2Rx8 - Dual Rank x8; JEDEC DDR4 standard 1.2V
- Improves system performance, workload capacity, and reduces bottlenecks by increasing memory (RAM) resources
- Note: This memory is ECC Unbuffered and cannot be mixed with different ECC types such as ECC Registered, ECC Load Reduced, or Non-ECC Unbuffered; (Memory compatibility can vary among different system models and their installed components; please verify compatibility and follow memory channel guidelines to ensure maximum performance)
Correct permissions and service working-directory settings
Linux services
Run the server as an unprivileged account that owns, or has explicit read and write access to, the server directory. Check that the account can create ShooterGame/Saved, append to the log, write crash output, and create downloaded mod files. A service can appear to start while silently failing when it cannot write its save or mod cache.
Set the service’s WorkingDirectory to the dedicated-server installation directory. The ARK dedicated-server documentation notes that a missing working directory can cause issues including failed automatic mod downloads. After changing it, stop the service, remove only incomplete temporary downloads if the log identifies them, and start once while watching the log for successful file creation.
Windows services
Check the account configured for the service or scheduled task. It must have modify permission on the installation and Saved directories. Confirm that User Account Control and security software are not blocking writes or quarantining server files. Test by launching under the same account used by the service; an administrator test under a different account can hide the real permission failure.
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Handle an engine hang safely
An engine hang leaves the process present, but the server stops accepting connections and CPU usage can fall to zero. Capture the current log, process state, and recent configuration or mod changes before terminating it. Check whether disk activity or memory is still changing; a short period of high disk activity during a save is different from a sustained zero-CPU stall.
Once evidence is preserved, an external restart policy can restore service availability. Configure a reasonable delay, write each restart to a separate service log, and alert after repeated failures. If every restart reaches the same map, mod, or memory event, disable the watchdog temporarily and return to controlled testing. A watchdog should never be the only response to a hang.
A repeatable recovery playbook
Use this decision path after the initial backup and evidence collection:
- Does the process exit? Check the crash dump and final log lines. If they indicate allocation failure, follow the memory procedure; if they name a mod or asset, perform the no-mod and file checks.
- Does it stay running with zero CPU and no connections? Treat it as a hang, preserve evidence, then restart once through a controlled policy.
- Does disabling all mods make it stable? Bisect the list, verify server-side files, and test each suspect on a copied save.
- Does only one save fail? Validate files, then compare fresh-world and backup results before touching the original.
- Does every map fail after an update? Validate the build, confirm client compatibility, and check for a service account or working-directory change.
- Does the log show write or download errors? Correct ownership, permissions, and
WorkingDirectorybefore changing content.
Prevent the next crash loop
- Schedule backups of
ShooterGame/Saved, retain multiple dated generations, and test restoring one to a separate directory. - Back up before every server update, mod removal, major configuration change, or map migration.
- Keep a change log with build, map, mod IDs, player count, and timestamps.
- Monitor RAM, swap or pagefile, CPU, disk capacity, process state, and restart count.
- Keep validation and rollback procedures documented for the account that operates the service.
- Use a small, intentional mod set and verify that each configured mod exists on the server.
- Set a service working directory explicitly and run with the least privilege that still permits required writes.
- Configure restart automation only after confirming that saves complete and logs survive a restart.
What to compare when choosing hosting or management
If the host itself is part of the problem, compare operational capabilities rather than headline CPU or RAM figures. Ask whether the service supports the ARK edition and map you run, how it handles mod installation and updates, whether backups are automatic and downloadable, how long logs are retained, whether RAM and disk graphs are available during a crash, and whether you can set a working directory or service account. Also check update scheduling, rollback controls, alerting, and restart limits. A panel that restarts quickly but discards logs is less useful than one that preserves evidence and lets you restore a known-good save.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWhen the server is stable again
After identifying the cause, change one variable at a time and let the server run through a normal save cycle, a map transition if you use one, and a representative player load. Keep the diagnostic backup until the world has passed that observation period. Record the fix, the evidence that confirmed it, and the rollback step so the next incident can be resolved without repeating destructive experiments.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
