Rent a managed server if you want a persistent Don’t Starve Together world without maintaining hardware, networking, updates and backups. Self-hosting is the better value when you already have suitable x86 hardware, understand SteamCMD and configuration files, and want complete control over mods, shards and ports. Renting costs a recurring fee but usually includes a control panel and some support; self-hosting can be inexpensive on existing equipment but makes you responsible for every failure and maintenance task.
The practical choice depends less on the game’s minimum specification than on who will operate the server. A private group with a technically confident owner can self-host successfully. A community that expects the world to remain online, needs caves and wants quick recovery will usually be better served by a managed host—provided the plan’s restart, backup, location and cancellation policies meet your needs.
Should you host a Don’t Starve Together server yourself or rent one?
Choose self-hosting for maximum control and the lowest possible ongoing cash cost when you already own an always-on computer and can maintain it. Choose renting for convenience, predictable availability and support. Neither option automatically delivers a better in-game experience: lag, downtime and world loss depend on the machine, network, configuration and recovery process you choose.
| Decision factor | Self-hosted server | Rented managed server |
|---|---|---|
| Administration | You install the server, apply updates, edit files, manage mods, monitor processes and make backups. | The provider supplies the machine and network environment; a panel and support reduce routine work. |
| Availability | Can remain online independently of a player’s game client, but only while your hardware, power and internet connection are working. | Designed for continuous availability; verify restart, maintenance and outage policies before subscribing. |
| Control | Direct access to cluster, shard, port and mod settings. | Controls are limited to what the provider exposes in its panel. |
| Cost model | Uses existing hardware or requires a purchase, plus electricity, maintenance and your time. | Has a visible recurring fee that may include backups, setup help and support. |
| Caves and scale | You run and resource the additional shard yourself. | Confirm that the selected plan supports both the surface and caves shards. |
| Recovery | You design the backup schedule, storage and restore procedure. | Compare the provider’s backup frequency, retention, restore method and geographic location. |
What does self-hosting a DST dedicated server involve?
Self-hosting means operating the dedicated-server software on your own Windows or Linux computer rather than running the world from a player’s client. The server process stores the world and accepts connections, so the owner does not need to be playing for other people to enter. That independence is the main reason groups self-host: the world can be available on a schedule chosen by the group, not by the person who happens to launch the game.
Hardware and operating-system baseline
Klei Entertainment’s 2025 dedicated-server documentation lists Windows or Linux, a 1.7 GHz or faster processor, at least 1 GB of RAM and at least 1 GB of hard-drive space. Klei also warns that the system should surpass the listed requirements. Those figures are a documented baseline, not a guarantee of smooth operation with caves, large worlds, many players or extensive mods.
| Requirement | Documented baseline | What it means in practice |
|---|---|---|
| Operating system | Windows or Linux | Use an operating system you can patch, secure and troubleshoot confidently. |
| Processor | 1.7+ GHz | Clock speed alone does not describe performance; leave headroom for the game server and the rest of the operating system. |
| Memory | 1+ GB RAM | The minimum may be adequate for a small, lightly modded setup, but caves and additional players increase demand. |
| Storage | 1+ GB hard-drive space | Reserve room beyond the installation for world data, logs, updates and multiple backups. |
| Connectivity | Broadband internet | Your upload capacity, latency, firewall and power reliability affect players more than the download speed shown by an internet plan. |
Software, tokens and shards
The dedicated-server software is distributed through SteamCMD or Steam. An online server also requires a Klei cluster token. The token identifies the cluster for online play; keep it private and treat it like a credential.
A cluster can contain multiple shards. The usual surface-and-caves arrangement uses one server instance for the surface and another for caves. Both processes, their configuration files and their network settings must be kept consistent. Adding caves therefore changes the resource and troubleshooting plan rather than being a cosmetic switch.
The ownership work people underestimate
Installation is only the first session. A self-hosted operator is responsible for:
- Installing or updating the server through SteamCMD or Steam.
- Creating the cluster and shard directories and keeping persistent world data in the intended locations.
- Choosing ports, forwarding them through the router when necessary and allowing the processes through the host firewall.
- Applying game updates and checking whether updates affect mods or configuration files.
- Adding, removing and ordering Workshop modifications, then testing that the world still loads.
- Watching CPU, memory, storage, logs and process status while the world is active.
- Making backups that are stored somewhere other than the live server and testing that a restore is possible.
- Explaining connection details to players and diagnosing failures caused by the home network, internet provider or local machine.
Klei’s command-line documentation covers persistent storage, cluster and shard paths, ports and runtime options, and notes that commands are used at the operator’s own risk. In other words, self-hosting transfers the provider’s operations job to the group owner.
What does renting a managed DST server provide?
A managed host supplies the physical or virtual machine, its network environment and a management interface. You generally select a plan, choose a location, configure the world in a panel and let the provider handle the underlying operating system. The exact boundary varies: some hosts provide only a basic control panel, while others add setup assistance, automatic backups, mod installation and support.
Example price and included features
BerryByte currently advertises Don’t Starve Together hosting from $6.95 per month on a 2026 page observation. That is one provider’s current starting offer, not a market-wide average or a promise that every location, player count or billing term costs the same. The advertised package includes setup help, automatic backups, caves and Workshop-mod support.
| Rental feature to verify | Why it matters | Questions to ask before paying |
|---|---|---|
| Server location | Distance between players and the machine affects latency. | Which regions are available, and can the location be changed later? |
| Restart behavior | A crashed process or planned maintenance can interrupt a world. | Does the service auto-restart, and how are players notified? |
| Backups | A backup is useful only if it is recent, retained and restorable. | How often are backups made, how long are they kept, and can I trigger or download one? |
| Caves support | Caves require a separate shard and additional resources. | Are surface and caves included in the selected plan, with no separate add-on? |
| Mods | Workshop content can change update and recovery behavior. | Can I install, disable and reorder the mods I need? |
| Support | Technical help is the main convenience premium over self-hosting. | What channels and response targets apply to setup, outages and restores? |
| Cancellation and data export | You need a practical way to leave without losing the world. | When does billing stop, and can the cluster and shard data be retrieved? |
Is self-hosting cheaper than renting?
Self-hosting is cheaper only when the value of your existing equipment and time is low enough. A spare computer that is already powered on may add little cash expense, but electricity, replacement parts, broadband limits and your maintenance time still have a cost. Buying a machine solely for the server can exceed several months or years of rental fees before the first player connects.
| Self-hosting cost | How to account for it |
|---|---|
| Hardware | Use the purchase price of a new machine, or assign a fair value to a computer that could otherwise be sold or used for another task. |
| Electricity | Estimate the machine’s average draw over the hours it is running and your local electricity rate. |
| Internet and networking | Include any upgrade needed for upload capacity, a static address or a router capable of the required forwarding and firewall rules. |
| Storage and backups | Budget for separate backup storage; a second copy on the same disk does not protect against disk failure. |
| Maintenance time | Value updates, mod troubleshooting, monitoring and recovery at the rate you would pay for technical work. |
Renting replaces those variable costs with a recurring bill. A low headline price may exclude premium locations, larger plans, extra storage or features such as automated backups. Compare the total monthly charge and the included limits rather than the starting number alone.
How do caves and multiple shards affect the choice?
Caves are a separate shard, normally running as its own server instance. Self-hosting therefore requires enough processor, memory, storage and operating-system headroom for both processes, plus the networking and monitoring needed to keep them connected to the same cluster.
On a rental service, caves are not guaranteed merely because the host supports DST. Check the plan description or ask support whether both shards are included, whether they share the advertised resource limit and whether backups cover the entire cluster. A plan that supports only a surface shard may be unsuitable for a group that relies on caves.
Shard separation also affects recovery. A backup that captures only one shard can leave the world inconsistent after a failure. Ask how the provider snapshots the cluster and whether a restore returns the surface and caves to a coordinated point in time.
How do you self-host a DST dedicated server?
Use this sequence as an operational checklist. The exact file names and launch arguments depend on the current Klei documentation and your operating system, so record every path and option you choose.
- Prepare the machine. Install a supported Windows or Linux system, apply security updates, create a dedicated service account if practical and ensure the machine has reliable power and broadband connectivity.
- Install the server software. Use SteamCMD or Steam to obtain the dedicated-server files. Keep the installation directory separate from persistent world data so updates do not overwrite the wrong files.
- Create the cluster identity. Obtain the required Klei cluster token for online hosting and store it with restricted permissions. Do not post the token in a public chat, script repository or support ticket.
- Set persistent storage. Create the cluster and shard paths described by Klei’s command-line documentation. Confirm that the account running the server can read and write those directories.
- Configure the surface shard. Set the world options, server name, ports and administrative settings in the cluster and shard configuration files. Keep a dated copy of the original configuration before making changes.
- Add caves if needed. Create the caves shard and its configuration, then allocate a separate server instance and the additional ports required by the cluster. Treat surface and caves as two processes that must be started, stopped and monitored.
- Open the network path. Allow the selected ports through the local firewall and, for a home connection, forward them on the router to the server machine. Test from outside your home network rather than relying only on a local connection.
- Launch and observe. Start the shard processes, watch their logs for configuration or port errors and connect with a test player. Confirm that the world saves and that caves transition correctly if enabled.
- Document administration. Write down the install path, persistent-data path, ports, token location, mod list, start and stop procedure and the person authorized to change settings.
- Automate maintenance carefully. Schedule backups and planned restarts, but apply game or mod updates when someone can test the world. Keep a rollback copy before a major update.
How should you choose a rental plan?
Start with the world you actually intend to run, not the smallest advertised package. A private, lightly modded world has a different requirement from a public community server with caves and an expectation of continuous access.
- Define the workload. List the expected number of simultaneous players, whether caves are mandatory, the mods you plan to use and the hours when the server must be online.
- Confirm the technical scope. Verify that the plan supports the game version, both shards when required, Workshop mods and the configuration controls your group needs.
- Check operational policies. Read the restart, maintenance, outage, backup retention, restore, cancellation and data-export terms before entering payment details.
- Choose a nearby region. Select a location that minimizes latency for most players; a cheaper distant region can feel worse than a slightly more expensive local one.
- Test support before committing. Ask a concrete pre-sales question about caves, backups or a mod. The clarity and speed of the answer indicate how useful support may be during an outage.
- Keep your own copy. Even with automatic backups, periodically download or otherwise preserve the cluster data if the provider permits it.
Which option is better for availability and recovery?
Renting generally wins when the group expects the world to be available while the owner is asleep, traveling or offline. The provider’s machine and network are designed for hosting, but “managed” does not guarantee zero downtime. Planned maintenance, resource limits and a failed restart can still interrupt play, so read the service policy.
Self-hosting can be equally available on a well-maintained machine with stable power and internet service. A home server is vulnerable to router reboots, power cuts, changing public addresses, upstream outages and a household decision to shut the computer down. If no one is watching the machine, a small failure can become a long outage.
Backups are a separate decision
Do not treat a backup checkbox as proof that recovery is safe. A useful backup has a known schedule, enough retention to cover delayed corruption, storage separate from the live world and a documented restore procedure. For a self-hosted server, periodically copy the cluster and shard data to another disk or system and perform a test restore. For a rented server, ask whether backups include every shard and whether you can initiate a restore without waiting for support.
What are the control and customization trade-offs?
Self-hosting gives direct access to the files and runtime options that define the cluster, shards, ports and mods. That is valuable when you need an unusual configuration, want to inspect logs or prefer to automate your own administration. It also means a mistaken edit can prevent a shard from starting, and you are responsible for knowing which settings are safe to change.
A rented panel hides infrastructure details and may make common tasks—starting, stopping, changing a world option or installing a Workshop mod—faster. The trade-off is a provider-defined interface. You may not have shell access, unrestricted file access, your preferred automation or the ability to change every port and runtime option. Confirm the panel’s limits before assuming that “full control” means the same thing as self-hosting.
Can an old PC run a dedicated DST server?
It may, if it runs Windows or Linux and exceeds Klei’s documented baseline, but the age of the machine is not enough to answer reliably. Check the processor, available memory, free storage, network connection, power reliability and whether the computer can remain on without disrupting other household use.
Test the exact workload you plan to run. Start with the intended world settings, add caves and enable the mods your group will use, then watch the server while several players connect. If the machine swaps memory, fills its disk, drops the network connection or competes with another workload, the minimum specification is not an adequate capacity plan.
A spare x86 mini PC can be a sensible self-hosting platform because it uses less power than a full desktop, but purchase only after comparing its price and expected electricity cost with several months of managed hosting. The machine still needs a backup destination and a network path that you can troubleshoot.
What commonly goes wrong, and how do you fix it?
Players cannot connect
- Confirm that the server processes are running and listening on the configured ports.
- Check the operating-system firewall and, for home hosting, the router’s forwarding rule.
- Test from a different network to distinguish a local-address problem from an internet-facing problem.
- Verify that the cluster token and server configuration belong to the same cluster.
The world starts but caves do not
- Confirm that the caves shard has its own configuration and process.
- Check that the required ports are not duplicated by another service.
- Review both shard logs for path, permission or mod errors.
- On a rental plan, verify that caves are enabled and included rather than assuming the surface package covers them.
An update breaks the server or mods
- Stop the server cleanly and preserve the current world and configuration before changing files.
- Review the mod list and disable the newest or recently changed item to isolate the conflict.
- Apply updates when someone can test connections and shard transitions.
- Use a known-good backup if the updated world fails to load.
The server repeatedly crashes
- Check logs, available memory, disk space and operating-system events.
- Reduce simultaneous workload or temporarily remove mods to separate capacity problems from configuration problems.
- Make sure both shard processes are not competing for the same port or data path.
- If rented, provide the host with timestamps, logs and the affected shard so support can investigate efficiently.
When is a hybrid approach sensible?
A group does not have to make one permanent choice. You can self-host a private world while the owner is available and rent a separate always-on world for a larger community. Another approach is to start with a managed host while the group learns its desired player count, caves usage and mod set, then move to owned hardware only when the operating workload and savings are clear.
If you change providers or move from rental to self-hosting, plan the data transition before cancelling service. Ask for a complete export of the cluster and every shard, identify the operating system and paths you will use, and test the copied world on the new machine before the old service expires. Keep the original data until players have confirmed that the replacement works.
Final decision: which should you choose?
Rent when uptime, convenience and recovery matter more than infrastructure control. A managed service is the safer default for a group that wants caves, automatic backups, setup assistance or a world that remains available without a technically confident owner on call.
Self-host when you already have suitable x86 hardware and are willing to operate it. You gain direct control over cluster, shard, port and mod settings and may reduce cash costs, but you also accept responsibility for SteamCMD, networking, updates, monitoring and backups. Before choosing, write down who will perform those jobs; if nobody owns them, the apparent savings are not worth the risk.
Frequently Asked Questions
Do I need to leave my game open for other players to use a dedicated server?
No. A dedicated server runs independently of an individual player’s client session. The machine hosting it must remain powered, connected and running the server processes, but the owner does not have to stay in the game.
Can I add caves later after starting with a surface-only world?
A cluster can include multiple shards, with caves operating as a separate instance. Adding caves requires creating and configuring that shard, assigning its runtime and network resources, and confirming that any rental plan supports the additional instance.
Are automatic host backups enough to protect my DST world?
Not automatically. Check how often snapshots run, how long they are retained, whether every shard is included and how restoration works. Keep an independent copy when the service allows it, because a backup you cannot retrieve or restore is not a practical recovery plan.
What information should I give support when a rented server fails?
Provide the server name or account identifier, the approximate failure time, the affected shard, recent configuration or mod changes and relevant log excerpts. Those details help support distinguish a process crash, port problem, resource limit or mod failure without exposing your private cluster token.
Is a low monthly price a reliable way to compare DST hosts?
No. A starting price may apply only to a particular region, billing term or resource tier. Compare the complete charge alongside caves support, backup retention, restart behavior, mod controls, support terms and data export before deciding.