Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Need a server to try this on? Antihost offers Linux VPS plans with NVMe storage, full root access and a 14-day free trial.
Running Ollama on a Linux VPS is a four-part job: install the Ollama binary, run it as a systemd service, place models on storage large enough for them, and keep the API private unless you deliberately add authenticated remote access. The commands below cover installation, model downloads, persistent storage, GPU and memory settings, updates, and recovery after common failures.
Ollama listens on 127.0.0.1:11434 by default. That local-only default is the safe starting point: applications on the VPS can call the API without a key, while remote clients should reach it through a restricted private network or an authenticated reverse proxy rather than an unprotected public port.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minute1. Check the VPS before installing
Ollama’s Linux installation supports the two architectures most commonly offered by VPS providers: amd64 (x86-64) and ARM64. Confirm the architecture, available memory, disks, and whether the provider exposes a usable GPU before downloading anything.
uname -m
free -h
lsblk -o NAME,SIZE,TYPE,MOUNTPOINT,FSTYPE
nprocx86_64normally means amd64;aarch64means ARM64.- Keep enough RAM for the operating system, Ollama, the model, and the context window. A model that technically fits in memory can still fail when several requests or models are active.
- Check persistent disk capacity separately from RAM. Model files remain after a reboot, so a small root disk can fill even when the VPS has plenty of memory.
- Confirm that the VPS permits systemd services. If the image uses a restricted container without systemd, use its supervisor instead of the unit shown below.
GPU acceleration is optional. The official Linux instructions document installation paths for NVIDIA CUDA and AMD ROCm. Without a supported GPU, Ollama can run on the CPU, provided the VPS has enough system memory and the expected performance is acceptable.
#1 Best Overall
- Powerful Turbo Fan:CAGIWIRU Electric Air Duster have reached speed up to 130000RPM,Can efficient remove dust and debris.With LED light and Three Adjustable Speed to meet with different cleaning tasks.
- With Attached USB C Fasting Charging cable.Charing 3 hours,Enjoy up to 240 minutes of use on the lowest setting, For third speed can use for 25mins.with four charging options to meet with your needs.
- Wide Application: attached 5 different nozzles,making it suitable for a variety of scenes, such as car,pet, pc, keyboards, and other electronic devices. It also serves for office and home clean duster.
- Handy Design & Portable:Cordless air duster electric nozzle is Light weight and small size, design and comfortable to hold and space saving, convenient to carry around. You can put it in your room or inside the car without taking up space.
- Economical &Reusable: Built durable materials with a long-lasting battery,our CAGIWIRU air duster it's a sustainable alternative to disposable air cans.
2. Install Ollama on Linux
Use the installer
On a supported Linux VPS, the shortest installation is:
curl -fsSL https://ollama.com/install.sh | shThe installer places the executable and, on distributions where the setup supports it, may create service-related files. Treat the downloaded script as a privileged installation step: inspect your provider’s image, run it on a maintained system, and verify the resulting binary and service before opening any network access.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsCheck that the command is available:
ollama --version
command -v ollamaInstall from an architecture-specific archive
If policy does not allow piping a remote script to a shell, download the archive for the VPS architecture from Ollama’s official Linux instructions, verify it according to the release instructions, and extract the binary into a directory on PATH such as /usr/local/bin. Do not use an amd64 archive on an ARM64 VPS, or the executable will fail before a model can be loaded.
The same Linux instructions describe optional NVIDIA CUDA and AMD ROCm packages. Install only the path matching the actual GPU and driver stack supplied by the VPS; a CPU-only machine does not need either package.
3. Run Ollama as a persistent systemd service
A service prevents the daemon from disappearing when your SSH session closes and starts it again after a reboot. Run these steps as a user with sudo access.
Create a dedicated account and model directory
sudo groupadd --system ollama
sudo useradd --system --gid ollama --home-dir /usr/share/ollama
--shell /usr/sbin/nologin ollama
sudo install -d -o ollama -g ollama /usr/share/ollamaIf the installer already created the ollama user or group, keep those existing objects and run only the directory command. The service should not run as root; a dedicated account limits the damage if the daemon or an extension is compromised.
Free tools Windows power users keep installed
One-click scans. No signup required.
Create the unit
Create /etc/systemd/system/ollama.service:
sudo tee /etc/systemd/system/ollama.service >/dev/null <<'EOF'
[Unit]
Description=Ollama Service
After=network-online.target
Wants=network-online.target
[Service]
ExecStart=/usr/bin/ollama serve
User=ollama
Group=ollama
Restart=always
RestartSec=3
[Install]
WantedBy=multi-user.target
EOFUse the actual path returned by command -v ollama in ExecStart. If it is /usr/local/bin/ollama, substitute that path. The unit deliberately binds to the default local address until you make an explicit networking decision.
Enable and start it
sudo systemctl daemon-reload
sudo systemctl enable ollama
sudo systemctl start ollama
sudo systemctl status ollamaA healthy status shows active (running). Test the local HTTP listener from the VPS itself:
Rank #2
- Screen Cleaner is Your Screen’s New Bestfriend! - EVEO is proud to present a screen cleaner that’s perfectly suited for your TV. Meet your new screen cleaner, computer screen cleaner, laptop cleaner, iPad cleaner, Macbook,tv cleaner or any other electronic device. Our screen cleaner spray and wipe along with the included plush microfiber cloth, easily removes dust, fingerprints & other smudges on the screen’s surface. Clean effectively for a streak-free screen
- Gentle on your Screens - Our screen cleaner spray and wipes is compatible to be used even for the most sensitive LCD, LED, CRT, and OLED screens in proper use. Be at ease knowing that this product is compatible with all the major brands: LG, SONY, Samsung, Sharp, iPad, iPhone, Android screens, Kindle, PC monitors - we’ve got you covered! Can be used as LCD screen cleaner, tv cleaner, smart tv screen cleaner, and more
- Premium Super Soft Plush Microfiber Cloth - Included in this efficient screen cleaning kit. This screen cleaner spray and wipe has our signature plush microfiber cloth specially designed to comply with screens and monitors and leave them streak-free. Has been shown to effectively absorbs stubborn stains, zero streaks, and no fiber shedding. Always use this together with our screen cleaner for the best results
- Easy-to-Use Screen Cleaner Kit + microfiber cleaning cloth - The EVEO Screen Cleaning Kit will easily be your new favorite tool when it comes to taking care of the things you love most, quick and easy solution for a streak-free screen. Simply screw on the sprayer gun on the screen cleaner bottle. Spray the cleaning liquid onto the microfiber cloth, then wipe the screen until desired cleanliness & shine is achieved.
- Quality Screen Care Kit Worth Purchasing - Enjoy a convenient Screen Cleaner anytime, anywhere you need it. Your satisfaction is our top priority. We stand behind the quality of our products and that’s why, you shouldn’t be worrying at all. We manufacture the Best Screen Cleaner. and can be used for laptop cleaning kit, laptop screen cleaner, tv screen cleaner for smart tv. MacBook screen cleaner, monitor cleaner or screen cleaner spray for electronic devices with microfiber cleaning cloth
curl http://127.0.0.1:11434/api/tagsAn empty model list is still a successful service response; it means the daemon is reachable but no model has been downloaded yet.
4. Download, run, inspect, and remove models
Ollama’s command-line workflow is intentionally small. Replace <model> with a model name and, where applicable, its tag.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →| Task | Command | What to expect |
|---|---|---|
| Download a model | ollama pull <model> |
Downloads the model layers into the configured model directory. |
| Start an interactive session | ollama run <model> |
Loads the model and opens a terminal chat session. |
| List installed models | ollama list |
Shows downloaded names, tags, sizes, and update information reported by Ollama. |
| Inspect metadata | ollama show <model> |
Displays model details, including the configuration Ollama knows for that tag. |
| See loaded models | ollama ps |
Shows models currently resident in memory and whether they are on CPU, GPU, or split. |
| Unload a model | ollama stop <model> |
Removes the model from active memory without deleting its files. |
| Delete a model | ollama rm <model> |
Removes the downloaded model data; pull it again before using it. |
For a first smoke test, pull a model that fits the machine, then run it:
ollama pull gemma4:e2b
ollama run gemma4:e2bThe Ollama quickstart describes this example download as approximately 7.2 GB and recommends 8 GB of available VRAM or unified memory. That is an example requirement, not a universal VPS specification: larger models and larger context windows need more memory.
When the daemon is managed by systemd, use the same commands as the account that can access the service. Model files belong to the service account, so a shell user without suitable permissions may see an empty inventory even though the service has models installed.
5. Move models to a larger disk
The documented Linux default model directory is:
/usr/share/ollama/.ollama/modelsOn a VPS with a small root volume, attach or mount a persistent disk, create a destination directory, copy the existing data, and then point the service at the new location with OLLAMA_MODELS.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Prepare and copy the destination
Assume the larger filesystem is mounted at /mnt/ollama. Choose a path on that filesystem and give the service account ownership:
sudo install -d -o ollama -g ollama /mnt/ollama/models
sudo rsync -a --info=progress2 /usr/share/ollama/.ollama/models/
/mnt/ollama/models/
sudo chown -R ollama:ollama /mnt/ollama/modelsStop the service during the final copy if models may change while rsync runs:
Rank #3
- 【Large size】：Measuring 800 x 300 mm/ 31. 5×12 inches, which is wider and taller than others. Large mouse pad can perfectly fit your keyboard and mouse of any size, and there is also ample space for peripherals such as phones, tablets, etc.
- 【LED backlight mouse pad】：Blade Hawks RGB mouse pad provides 7 static modes and 7 dynamic modes will give you the options you are looking for. Power-off memory function remembers the last lighting mode you selected. The super glow fiber Chroma will give you a colorful gaming setup that you want.
- 【Anti-slip rubber base】：Made from Natural rubber, this gaming mousepad will firmly grip your desktop, allowing you to enjoy the ultimate precision in the games you are most passionate about. It also comes with a texture that enhances this ability even further.
- 【Smooth and waterproof surface】: Micro-textured cloth surface, for extremely precise mouse control and a smooth movement. When liquid splashes on the gaming mouse pad, it forms water droplets and slides down. Will not delay your work or gameplay.
- 【Plug and play】: This mouse and keyboard pad is powered by USB, plug and play, no driver required. One button control light modes. Press one time to change the lighting mode, press twice to change the brightness, press and hold about 3 seconds turn ON/OFF.
sudo systemctl stop ollama
sudo rsync -a --delete /usr/share/ollama/.ollama/models/
/mnt/ollama/models/
sudo chown -R ollama:ollama /mnt/ollama/modelsSet OLLAMA_MODELS in a systemd drop-in
Use the systemd-supported override rather than editing the vendor unit directly:
sudo systemctl edit ollama.serviceAdd:
[Service]
Environment="OLLAMA_MODELS=/mnt/ollama/models"Save the override, then reload and restart:
sudo systemctl daemon-reload
sudo systemctl restart ollama
sudo systemctl status ollamaConfirm that the service can see the inventory:
ollama list
ollama pull <small-test-model>Do not delete the old directory until the service has restarted successfully and the model list is correct. If the list is empty, check the drop-in, the mount point, and ownership first:
systemctl cat ollama.service
sudo -u ollama test -r /mnt/ollama/models
sudo -u ollama test -w /mnt/ollama/models
findmnt /mnt/ollamaThe destination must be mounted at boot. Otherwise systemd can start Ollama against an empty directory before the disk is available, making the models appear to have vanished.
6. Configure memory, context, and concurrency
Ollama exposes service environment variables for common deployment controls. Put persistent values in the same systemd drop-in used for storage, beneath one [Service] section.
| Variable | Purpose | Documented default or behavior |
|---|---|---|
OLLAMA_HOST |
Bind address and port | Local binding is 127.0.0.1:11434 by default. |
OLLAMA_CONTEXT_LENGTH |
Default context window | 4096 tokens; can also be overridden per request with num_ctx. |
OLLAMA_KEEP_ALIVE |
How long a model remains resident | Controls residency so repeated requests do not always reload the model. |
OLLAMA_MAX_LOADED_MODELS |
Maximum simultaneously loaded models | Defaults to three times the number of GPUs, or three for CPU inference. |
OLLAMA_NUM_PARALLEL |
Parallel requests per model | Defaults to one. |
OLLAMA_MAX_QUEUE |
Requests waiting for capacity | Defaults to 512. |
For example, a conservative CPU or single-GPU service might use:
[Service]
Environment="OLLAMA_MODELS=/mnt/ollama/models"
Environment="OLLAMA_CONTEXT_LENGTH=4096"
Environment="OLLAMA_NUM_PARALLEL=1"
Environment="OLLAMA_MAX_QUEUE=64"After changing values:
sudo systemctl daemon-reload
sudo systemctl restart ollamaContext length and parallelism have direct memory consequences. Parallel processing increases memory use in proportion to context length. Start with one parallel request, observe real workloads, and increase limits only when ollama ps, system memory, GPU memory, and service logs show headroom.
Per-request context
The API can override the default context for a particular request with the num_ctx option. A larger value helps long prompts but consumes more memory, so do not raise it globally merely because one workflow needs a long document.
7. Verify CPU and GPU placement
Run:
ollama psThe processor column reports values such as 100% GPU, 100% CPU, or a split such as 48%/52% CPU/GPU. A split means the model is too large, or otherwise configured, to reside entirely in one accelerator and has been distributed across available resources.
Rank #4
- 【Softer and More Comfortable】Vaydeer wrist rest has unique diamond pattern, which is the combination of softness and aesthetics. The materials of wrist rest are improved into higher quality memory foam and covered with silky smooth lycra. The computer wrist rest makes you as comfortable and cushiony as like rest your wrists on clouds.
- 【Ergonomic Wrist Saver】The wrist rests for keyboard and mouse comes with a 17.32×3.15×0.83 inch keyboard wrist pad and a 5.94×3.15×0.83 inch mouse wrist support. Based on ergonomic design, the unique concave shape is the perfect fit for your wrist joints. The wrist rest pad fits most computer keyboards and laptops, improve hand and wrist posture, release your wrist and arm stress.
- 【Non-Slip Rubber Bottom】Featuring an anti-skid silicone base on the bottom, this wrist keyboard support stays firmly in place on your desk, preventing the padding from sliding around, ensuring stable and consistent wrist support during extended computer sessions.
- 【Better Experience & Pain Relief】Our keyboard arm rest is beneficial to alleviate the soreness caused by direct contact and friction between your arm and a hard desk surface, reducing the risk of wrist fatigue or carpal tunnel. The soft texture of memory foam can evenly distribute the pressure around your wrists and provide good support with just enough give.
- 【Helpful in Multiple Scenarios】Whether you're working, studying, writing, typing, gaming, this keyboard and mouse rest combo is an essential accessory to add comfort and support to your hands and wrists. It’s also a great gift for men, women, family, friend, coworker, gamer, teacher, etc.
The FAQ describes a memory-driven placement rule: if a model fits on one GPU, Ollama loads it there; if it does not, Ollama may spread it across available GPUs. Loading several models and accepting parallel requests consumes additional VRAM or RAM. A model that fits when idle can therefore be rejected later under concurrent traffic.
- For GPU-backed VPSs, compare usable VRAM, not just the GPU name.
- Reserve system RAM for the operating system and any CPU-offloaded layers.
- Use
ollama stop <model>to free model memory before testing another large model. - Keep context and parallel request settings modest until placement is stable.
8. Call the local API
The local API is available at http://localhost:11434. Local requests do not require an API key. A minimal chat request is:
curl http://127.0.0.1:11434/api/chat
-H 'Content-Type: application/json'
-d '{
"model": "gemma4:e2b",
"messages": [
{"role": "user", "content": "Reply with one short sentence."}
],
"stream": false
}'Use a model that is actually present in ollama list. If the request returns a model-not-found error, pull the exact name and tag first. If it hangs, inspect whether the model is downloading or loading and check the service log.
Applications running on the same VPS should call 127.0.0.1:11434 or localhost:11434. Keeping the listener local avoids exposing the unauthenticated local interface to the internet.
9. Expose the API safely when remote access is required
Change the bind address only when a remote client genuinely needs direct network reachability. Set OLLAMA_HOST in the systemd drop-in, for example:
[Service]
Environment="OLLAMA_HOST=0.0.0.0:11434"Then reload and restart the service. Binding to all interfaces makes the port reachable anywhere allowed by the VPS firewall and provider network; it does not add authentication by itself.
The safer production pattern is to keep Ollama on a private interface and put an access-controlled reverse proxy or private tunnel in front of it. Apply all of the following:
- Require authentication at the proxy or identity layer. The local API’s no-key behavior is not a substitute for remote access control.
- Allow only the required source addresses or private-network peers in the firewall.
- Terminate TLS at the proxy when traffic crosses an untrusted network.
- Do not publish port 11434 directly through a broad “allow from anywhere” firewall rule.
- Log and rate-limit requests, especially if the endpoint can trigger large model loads or long generations.
Ollama’s FAQ documents reverse-proxy use with Nginx and tunnel approaches. Whatever front end you choose, verify it forwards the request body and streaming responses correctly, and test from an approved remote client while confirming that unapproved sources are rejected.
Best Value
- 【No Drill Mounted】Cable management tray can be assembled in as little as 3 minutes or less-simply. Side clips easily clampe the edge of desk, pass through the holes on the sides, and secure excess cables with baskets. Desk cable management supports inward or outward installation to meet your needs in different scenarios, and provides you with great convenience when collecting and organizing wires.
- 【Rubber Pad Protects Desk】Wire organizer under desk built-in soft rubber pads on the clamp protect your desk from scratches, so you will not damage your furniture or office.
- 【Sturdy and Beautiful】Cord management under desk is made of high-quality carbon steel and holds about 15 lbs. The wire basket has 2 holes on each side of the tray to enable your cables to pass freely. Cable baskets are suitable for desk thickness use from 0.4" to 2.4".
- 【Space Saving & Keep Organized】Desk cable management tray holds all your power cords, outlet strips, USB hubs and computer transformers hidden from your feet, and the data cables will not enter the vacuum cleaner to keep your desk clean and tidy. Desk wire organizer is also suitable for use in the kitchen and outdoors.
- 【Reduce Safety Risks】Cable tray under desk keeps all plugs away from your kids, no more worry about tripping. Using 2 holes wire tray saves extra space, helps you easily cabling and wire protection. The mesh design is not easy to accumulate dust, but avoids safety accidents caused by messy wires.
10. Update Ollama and pin a release
To update an installation made with the Linux installer, rerun the installer. For an archive installation, download the newer archive and replace the binary using the documented procedure. Pin a specific installer release by setting OLLAMA_VERSION, for example:
Quick wins for a faster PC:
Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →OLLAMA_VERSION=0.5.7 curl -fsSL https://ollama.com/install.sh | shUse a version pin when you need repeatable builds or want to stage an upgrade. Before replacing a production binary, record the current version and test the service on a separate VPS or maintenance window.
After an update:
sudo systemctl daemon-reload
sudo systemctl restart ollama
sudo systemctl status ollama
ollama --versionModel data normally remains in the configured OLLAMA_MODELS directory, but keep a backup or reproducible pull plan for models that are costly to redownload.
11. Read logs and troubleshoot failures
Use the journal first:
journalctl -e -u ollama
journalctl -u ollama -f| Symptom | Likely checks | Action |
|---|---|---|
systemctl start fails |
Wrong binary path, missing user, malformed unit | Run systemctl cat ollama.service, verify ExecStart, and inspect the journal for the first error. |
| API connection refused | Service stopped or listening elsewhere | Check systemctl status ollama, the configured OLLAMA_HOST, and local connectivity with curl. |
| Models disappeared after moving storage | Drop-in not loaded, mount missing, permissions wrong | Check systemctl cat, findmnt, and read/write access as the ollama user. |
| Model will not load | Insufficient VRAM/RAM or an oversized context | Stop another model, reduce context or parallelism, and review ollama ps and logs. |
| Requests remain queued | All loaded capacity is busy | Keep parallelism conservative, confirm memory headroom, and raise queue limits only with a measured workload. |
| Remote client cannot connect | Local-only bind, firewall, proxy, or private-route issue | Test locally first, then verify the bind address and each network control without exposing the raw port broadly. |
Permission errors after changing the model directory
Every directory component must be traversable by the ollama user, and model files must be readable and writable where Ollama needs to update them. Check:
namei -l /mnt/ollama/models
sudo -u ollama ls -la /mnt/ollama/models
sudo -u ollama sh -c 'touch /mnt/ollama/models/.write-test'
sudo -u ollama rm /mnt/ollama/models/.write-testCorrect ownership on the model tree and ensure the disk is mounted before the service starts. Avoid solving a permission problem by running the daemon as root.
12. Choose a VPS by workload, not by a single model label
Ollama’s official documentation does not establish one universal VPS size. The right choice depends on model memory, context length, concurrency, storage growth, and network design. Use these criteria when comparing providers or plans:
| Criterion | Questions to answer | Why it matters |
|---|---|---|
| GPU type and VRAM | How much usable VRAM is available, and is the driver stack compatible? | Determines whether a model fits fully on the GPU or spills to system memory. |
| CPU and system RAM | How much RAM remains after the OS and other services? | CPU inference and GPU-offloaded layers rely on system memory; concurrency adds demand. |
| Persistent disk and I/O | Can the disk hold current and future model files, and is throughput adequate? | Pulls, startup loads, updates, and storage migrations are I/O-heavy. |
| Architecture | Is the machine amd64 or ARM64, and is the required Ollama build available? | The installer or archive must match the CPU architecture. |
| Network egress and private networking | Are private subnets, firewall rules, and predictable egress available? | Remote API access should be restricted rather than exposing port 11434 publicly. |
| Service control | Can you run systemd or an equivalent supervisor and mount disks at boot? | Persistent operation and reliable model storage depend on both. |
| Region and availability | Is the required GPU capacity available in the reader’s region, and at what recurring price? | GPU inventory and pricing vary by geography and can change independently of Ollama. |
For a small internal assistant, a CPU VPS with adequate RAM and a large persistent disk may be sufficient. For larger models, long contexts, or several simultaneous users, GPU VRAM and memory bandwidth become the limiting factors. Size from the measured model footprint and expected concurrency rather than from a provider’s marketing tier name.
Quick Recap
13. A repeatable deployment checklist
- Confirm
amd64orARM64, RAM, disk, GPU, and systemd availability. - Install Ollama with the Linux installer or the matching architecture archive.
- Create the unprivileged
ollamaaccount and verify the binary path. - Create and enable the systemd unit; confirm
active (running). - Call
http://127.0.0.1:11434/api/tagslocally. - Pull one model, run it interactively, and inspect it with
ollama ps. - Move models to a mounted disk before the root filesystem fills, using
OLLAMA_MODELSin a drop-in. - Set context, keep-alive, loaded-model, parallel, and queue limits conservatively.
- Keep
OLLAMA_HOSTlocal unless remote access is required. - If remote access is required, place an authenticated, restricted proxy or private network in front of Ollama.
- Record the version, test updates, and monitor
journalctl -u ollamaduring real workloads.
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.
