Self-Hosted
/
Docker-OSX
mirror of https://github.com/sickcodes/Docker-OSX.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from '9fead74c71'
${ noResults }
Docker-OSX/helm
History
peter 2567054805
remove cmd to install kernel in pod runtime as it has been migrated to Dockerfile 		4 years ago
..
templates remove cmd to install kernel in pod runtime as it has been migrated to Dockerfile 4 years ago
.DS_Store
…
Chart.yaml update doc with features 4 years ago
Dockerfile add gpu passthrough support as well as dynamic OpenCore regeneration 4 years ago
INSTALL-QEMU-AND-GPU-IOMMU.md add gpu passthrough support as well as dynamic OpenCore regeneration 4 years ago
README.md add gpu passthrough support as well as dynamic OpenCore regeneration 4 years ago
values.yaml remove cmd to install kernel in pod runtime as it has been migrated to Dockerfile 4 years ago

README.md

docker-osx

Information

This installs docker-osx in Kubernetes.

Features

What works

  1. Setting cpu/memory options
  2. Setting VNC password
  3. Persistance
  4. Setting SMBIOS
  5. QEMU/virtio cpu/software gpu changes
  6. Toggling Audio
  7. Additional port forwarding
  8. Kubernetes resource requests/limits
  9. Defining version of macOS to install
  10. Defining install partition size
  11. Defining a different version of macOS
  12. Additional QEMU parameters
  13. GPU support

What doesn't

  1. Simultaneous VNC + GPU IOMMU (this is a limitation of QEMU :( unfortunately)

Requirements

*) Install host machine requirements *) Ensure you are running QEMU 5.X *) Kubernetes *) Helm v2 *) sickcodes/docker-osx-vnc Docker image

Build sickcodes/docker-osx-vnc

  1. Go back to the root directory

  2. Build docker image

    docker build \
    -t sickcodes/docker-osx-vnc:latest \
    -f helm/Dockerfile .

    Please ensure you are using the Dockerfile in the helm folder

Do not worry about passing CPU, RAM, etc as they are handled in values.yaml now.

Installation

If planning on using a GPU with IOMMU passthrough it is recommended to configure it first and install macOS--otherwise installing may take a very long time depending on your hardware. Please see qemu.systemInstaller.downloadDelay, qemu.systemInstaller.cache, qemu.systemDisk.downloadDelay, and qemu.systemDisk.cache for possibly reducing installation time. It has taken me over three hours to install on some occasions with a NVMe secondary disk without GPU passthrough configured..

In values.yaml..

  1. Set a unique password for vnc.password.
  2. Re-generate SMBIOS configPlist.MLB, configPlist.SystemSerialNumber, and configPlist.SystemUUID for iServices to work.
  3. Update serverName to reflect the unique name (in the case more than one deployment is required).
  4. Configure qemu.systemInstaller.downloadDelay (in a period of seconds) that reflects how long your internet connection will download around 500MB (BaseSystem.dmg) + uncompress the file (which took about the same time for me to download on a 1gig internet connection).
  5. Set service.ip to reflect an IP address of your choice, or use ingress.
  6. Update extraVolumes.hostPath.path to something useful for you.

Optionally..

  1. Install kexts to kexts.path and enable.
  2. Adjust openCore.boot.timeout if desire for macOS to load automatically.
  3. Add usb devices with qemu.usb or qemu.extraArgs if desired.
  4. Add more ports for portforwarding services if needed.

Afterwards..

  1. Launch your VNC viewer of choice and connect to the IP/hostname you defined + the port 8888 with the password specified for vnc.password.
  2. Install macOS like usual.

Please note, after you have installed macOS feel free to set qemu.systemInstaller.downloadDelay to nothing, as BaseSystem.dmg will be stored in the path defined for extraVolumes.hostPath.path

Resources

Please note, resource limits may vary based on hardware. The ones currently defined are ones that worked for me personally.