Lima
Lima 2.2.0 provides lightweight Linux virtual machines with automatic file sharing and port forwarding, plus templates for containerd, Docker, Kubernetes and other isolated development environments.
What Lima does
Lima launches Linux virtual machines for development and testing, while helping configure host file sharing and port forwarding. Templates can provide containerd, Docker, Podman or Kubernetes workflows, but each runtime still has its own client socket, image store and lifecycle.
Version and platform boundaries
Version 2.2.0 includes macOS, Linux and Windows builds across Intel, AMD64, Apple silicon and ARM64 variants, along with guest agent assets. Match the limactl binary, guest components and host architecture instead of mixing files from different releases.
Instances, mounts and data
An instance contains a system disk, configuration, containers, images and volumes. Stopping an instance releases runtime resources; deleting it can remove data stored inside the virtual machine. Export images, databases, volumes and important files before destructive lifecycle operations.
Mount and network boundaries
Automatic mounts and forwarded ports are convenient but expose host files or services to the guest. Use the smallest dedicated directory, explicit read-only settings where possible and clearly named ports and client contexts.
Maintenance note
Review date: 2026-08-23. The page was checked against Lima 2.2.0, current platform assets, templates, guest agents, mount permissions and instance recovery boundaries.
Save to your cloud drive
Open the cloud drive to get the file directly, or save it for convenient access on another device.
Quark Cloud Drive
RecommendedSave Lima to this cloud drive
Baidu Netdisk
Save Lima to this cloud drive
Lima 2.2.0 installation, VM templates and container guide
Start with a disposable default instance, check Linux commands, port forwarding and a dedicated mount, then create separate named instances for container runtimes and export data before deletion.
Before you start
- Prepare hardware virtualization, disk space for Linux images and container layers, and a host supported by the selected build.
- Record existing instance names, templates, CPU, memory, disk, network and mount paths to avoid conflicts with Colima or Rancher Desktop.
- Prepare Docker CLI, nerdctl or kubectl only when the selected runtime needs it, and back up databases and named volumes before experiments.
Installation steps
- 01
Install the matching host and architecture package
Select the Intel, x86_64, Apple silicon, aarch64, AMD64 or ARM64 build for the host, unpack it into a trusted directory and verify limactl reports 2.2.0.
- 02
Create a disposable default instance
Run the default start flow, read the resources and mounts that will be created, then enter the instance and check the Linux version and network.
- 03
Use named templates for runtimes
Create a separate named instance for Docker, containerd or Kubernetes, test its client socket and configuration, and avoid overwriting an existing runtime context.
Quick start
- 01
Verify the instance lifecycle
Create a test file, list the instance, stop it, start it again and re-enter it. Confirm that the name and state you inspect match the intended instance.
- 02
Verify mounts and port forwarding
Run a small service on a test port and access it from the host, while mounting only a dedicated directory and checking read, write, symlink, case and performance behavior.
- 03
Verify a container runtime
Run a disposable container in the selected template, confirm the client is connected to the intended instance and check image and volume state after a restart.
Usage tips
- Mounted directories can expose real host data to guest processes. Keep mounts minimal and prefer read-only access for inspection tasks.
- Docker, Podman, containerd and Kubernetes use different templates, sockets and client contexts; name each one explicitly.
- A stopped instance is not a backup. Export the system data and validate a restore before removing an instance.
Troubleshooting and uninstall
Why is an instance stuck while preparing an image or network?
Check virtualization support, host version, disk space, proxy, DNS and image download status. Review the target instance log and confirm that no incompatible old configuration or duplicate instance is being used.
Why are mount permissions wrong or host files changed unexpectedly?
Stop the affected instance, inspect the template mount path, read-only flag and user mapping, then restore from an independent backup. Do not mount an entire home or secrets directory for an untrusted workload.
- Export instance dataStop writes, export images, databases and volumes, copy files that exist only in the VM and verify a restore before deleting the instance.
- Remove instances and LimaConfirm each instance name before deletion, remove the program and guest agent only after all retained data and templates have been archived.
Frequently asked questions
Is Lima only for running containers?
No. Lima provides template-driven Linux virtual machines with file sharing and port forwarding. Containers are a common workflow, but a VM can also serve ordinary development and testing needs.
How is Lima different from Colima?
Lima offers lower-level, template-based VM control. Colima builds on Lima and emphasizes a simpler command flow for Docker, containerd, Incus and Kubernetes environments.
Will deleting a Lima instance preserve its images and volumes?
Treat deletion as destructive for the system disk, images, volumes and configuration. Export the data and verify recovery before removing the instance.