Install Docker on Termux

answerfile

KEYMAPOPTS="us us"
HOSTNAMEOPTS="-n alpine"
INTERFACESOPTS="auto lo
iface lo inet loopback

auto eth0
iface eth0 inet dhcp
    hostname alpine
"
TIMEZONEOPTS="-z UTC"
PROXYOPTS="none"
APKREPOSOPTS="http://dl-cdn.alpinelinux.org/alpine/v3.12/main http://dl-cdn.alpinelinux.org/alpine/v3.12/community"
SSHDOPTS="-c openssh"
NTPOPTS="-c busybox"
DISKOPTS="-v -m sys -s 0 /dev/sda"

docker-termux.md

Docker on Termux [in a VM]

Create a Linux VM and install Docker in it so you can (slowly) run x86 Docker containers on your Android device.

Recommended to use SSH or external keyboard to execute the following commands unless you want sore thumbs. See https://wiki.termux.com/wiki/Remote_Access#SSH

• Install QEMU

   pkg install qemu-utils qemu-common qemu-system-x86_64-headless
  

• Download Alpine Linux 3.12 (virt optimized) ISO

   mkdir alpine && cd $_
   wget http://dl-cdn.alpinelinux.org/alpine/v3.12/releases/x86_64/alpine-virt-3.12.3-x86_64.iso
  

• Create disk (note it won't actually take 4GB of space, more like 500MB)

   qemu-img create -f qcow2 alpine.img 4G
  

• Boot it up

  qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64 \
    -drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-x86_64-code.fd \
    -netdev user,id=n1,hostfwd=tcp::2222-:22 -device virtio-net,netdev=n1 \
    -cdrom alpine-virt-3.12.3-x86_64.iso \
    -nographic alpine.img
  

• Login with user root (no password)

• Setup network (press Enter to use defaults):

   localhost:~# setup-interfaces
   Available interfaces are: eth0.
   Enter '?' for help on bridges, bonding and vlans.
   Which one do you want to initialize? (or '?' or 'done') [eth0] 
   Ip address for eth0? (or 'dhcp', 'none', '?') [dhcp] 
   Do you want to do any manual network configuration? [no] 
   localhost:~# ifup eth0
  

• Create an answerfile to speed up installation:

  localhost:~# wget https://gist.githubusercontent.com/oofnikj/e79aef095cd08756f7f26ed244355d62/raw/answerfile
  

• Patch setup-disk to enable serial console output on boot

  localhost:~# sed -i -E 's/(local kernel_opts)=.*/\1="console=ttyS0"/' /sbin/setup-disk
  

• Run setup to install to disk

  localhost:~# setup-alpine -f answerfile
  

• Once installation is complete, power off the VM (command poweroff) and boot again without cdrom:

  qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64 \
    -drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-x86_64-code.fd \
    -netdev user,id=n1,hostfwd=tcp::2222-:22 -device virtio-net,netdev=n1 \
    -nographic alpine.img
  

• Install docker and enable on boot:

  alpine:~# apk update && apk add docker
  alpine:~# service docker start
  alpine:~# rc-update add docker
  

• Useful keys:

  • Ctrl+a x: quit emulation
  • Ctrl+a h: toggle QEMU console

Perhaps the issue might be located within the QEMU starting command itself. You need to specify the hostfwd flag along with the port you intend to use. For example, use -hostfwd=tcp::9000-:9001. What this flag does is to forward TCP traffic from port 9000 on the host to port 9001 on the guest. In other words, any incoming traffic directed at port 9000 on the host will be redirected to port 9001 within the virtual machine.

I don’t know if it is right because if I have more services with different ports?

I don't know if there's a better way to do this, but if you want to forward multiple ports, you could've use the hostfwd flag multiple times. This is how the command would looks like -hostfwd=tcp::8080-:80,hostfwd=tcp::9000-:8081. This is how i usually forward ports on QEMU so, hope it helps.

-hostfwd=tcp::8080-:80 -hostfwd=tcp::9000-:8081

if I use this command:
qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64 -drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-x86_64-code.fd -netdev user,id=n1, -hostfwd=tcp::2222-:22 -hostfwd=tcp::8080-:80 -hostfwd=tcp::9000-:8081 -device virtio-net,netdev=n1 -nographic alpine.img

I gets this message:
qemu-system-x86_64: -hostfwd=tcp::2222-:22: invalid option

@LeftandRights maybe I think it looks like that portforwarding between Termux and VM Alpine not working because If I try ping in termux to vm alpine (ping 10.0.2.15) I get icmp_seq=10 Time to live exceeded, but in VM Alpine is eth0: 10.0.2.15. What do you mean?

@LeftandRights maybe I think it looks like that portforwarding between Termux and VM Alpine not working because If I try ping in termux to vm alpine (ping 10.0.2.15) I get icmp_seq=10 Time to live exceeded, but in VM Alpine is eth0: 10.0.2.15. What do you mean?

Although it shows 10.0.2.15 on the VM, it actually forwarded to your machine IPv4 address (192.168.1.32)

@LeftandRights maybe I think it looks like that portforwarding between Termux and VM Alpine not working because If I try ping in termux to vm alpine (ping 10.0.2.15) I get icmp_seq=10 Time to live exceeded, but in VM Alpine is eth0: 10.0.2.15. What do you mean?

Although it shows 10.0.2.15 on the VM, it actually forwarded to your machine IPv4 address (192.168.1.32)

OK, if I try ping from VM Alpine to Termux so in alpine:~# ping 192.168.1.32 I get: 64 bytes from 192.168.1.32: seq=2 ttl=255 time=6.812 ms. In this case it looks ok

@LeftandRights so this mean that I can from VM Alpine to Termux but I cann't from Termux to VM Alpine?

-hostfwd=tcp::8080-:80 -hostfwd=tcp::9000-:8081

if I use this command: qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64 -drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-x86_64-code.fd -netdev user,id=n1, -hostfwd=tcp::2222-:22 -hostfwd=tcp::8080-:80 -hostfwd=tcp::9000-:8081 -device virtio-net,netdev=n1 -nographic alpine.img

I gets this message: qemu-system-x86_64: -hostfwd=tcp::2222-:22: invalid option

Sorry for the previous comment, the hostfwd flag should be used along with the netdev flag which turns out be something like this
qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64 -drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-x86_64-code.fd -netdev user,id=n1,hostfwd=tcp::2222-:22,hostfwd=tcp::8080-:80,hostfwd=tcp::9000-:8081 -device virtio-net,netdev=n1 -nographic alpine.img

When I launch Alpine with QEMU, I obtain this :

" Starting networking ... * lo ... [ ok ]

• docker0 ...Cannot find device "docker0"
  Cannot find device "docker0"
  Device "docker0" does not exist.
  [ ok ]
• eth0 ... [ ok ]
• Seeding random number generator ... * Seeding 256 bits and crediting
• Saving 256 bits of creditable seed for next boot
  [ ok ] "

got internet error. so what I did:
vi /etc/resolv.conf

nameserver 8.8.8.8

:wq

To avoid the connection problem that occurred with the authors of the posts above, you need to do the following:

1. In the interface setup item localhost:~# setup-interfaces, select 2 items and 3 type "yes" instead of [no]
   Do you want to do any manual network configuration? [no] yes
2. In the document that opens, paste this:
   auto lo
iface lo inet loopback
auto eth0
iface eth0 inet static
address 192.168.1.10
netmask 255.255.255.0
gateway 192.168.1.1
dns-nameservers 1.1.1.1
3. Check:
   cat /etc/network/interfaces
4. localhost:~# ifup eth0
5. Check this file:
   cat /etc/resolve.conf
6. If there is no such thing, then insert:
   echo "nameserver 192.168.1.1 nameserver 1.1.1.1" > /etc/resolv.conf
7. localhost:~# /etc/init.d/networking restart
8. if there are no errors, everything is ok

echo "nameserver 192.168.1.1
nameserver 1.1.1.1" > /etc/resolv.conf

This works

Hello! 👋

I forked @egandro's qemu-docker-arm (great work!), did some refactoring to work with the latest version of Alpine (3.19) and added to the README @schnatterer's idea to connect and use docker in your terminal using ssh! Creating a great experience to use docker without needing root! 😄

repo: https://github.com/AntonyZ89/docker-qemu-arm

using docker compose to up my postgresql + redis:

Screen_Recording_20240118_144445_Termux-ezgif.com-crop-video.mp4

Thank you good sir !

i have no internet why?

localhost:# ifup eth0 udhcpc: started, v1.31.1 udhcpc: sending discover udhcpc: sending select for 10.0.2.15 udhcpc: lease of 10.0.2.15 obtained, lease time 86400 localhost:# wget https://gist.githubusercontent.com/oofnikj/e79aef095cd08756f7f 26ed244355d62/raw/answerfile wget: bad address 'gist.githubusercontent.com' edit: This network setting work for me: localhost: # echo -e "auto lo\niface lo inet loopback\nauto eth0\niface eth0 inet dhcp" > /etc/network/interfaces localhost: # /etc/init.d/networking restart localhost:~# echo nameserver 1.1.1.1 > /etc/resolv.conf nameserver 8.8.8.8 does not work for me

Thanks for solving my problem. For me other problem is need modify /etc/apk/repositories enable community source, delete # in '#http://dl-cdn.alpinelinux.org/alpine/v3.12/community', otherwise run 'apk add docker' with error 'required by: world[docker-compose]'.

I tried running a Flask server in it and listen to port 5000 and host 0.0.0.0 which supposed to be accessible to all devices that connected to the same router. But in this case i can't even access it through my browser. Feel free to reply, any help will be appreciated.

qemu-system-x86_64 -machine q35 -m 1024 -smp cpus=2 -cpu qemu64
-drive if=pflash,format=raw,read-only=on,file=$PREFIX/share/qemu/edk2-x86_64-code.fd
-netdev user,id=n1,hostfwd=tcp::2222-:22,hostfwd=tcp::9000-:9000,hostfwd=tcp::4567-:4567,hostfwd=tcp::5344-:5344 -device virtio-net,netdev=n1
-cdrom alpine-virt-3.12.3-x86_64.iso -nographic alpine.img

like this ,portainer is working now

This is crazy!! Spinning up a vm in my mobile and running docker in it 🚀

I updated my installer: https://github.com/egandro/docker-qemu-arm

• latest (stable) alpine 3.19
• that means latest docker + docker compose
• mitigation for (well known) alpine dns and ipv4/v6 issues

Please report issues here: https://github.com/egandro/docker-qemu-arm/issues

Thanks for updated to 3.19.1

I am trying to install with the aarch64 qemu but keeps getting
/dev/sda is not a block device suitable for partitioning.

All the step went fine but getting error at the Disk & Install step, looks like it is not able to find the disk.
which I created using:
qemu-img create -f qcow2 alpine.img 20G

I am using below command to boot:
qemu-system-aarch64 -machine virt -m 2048 -smp cpus=2 -cpu max
-drive if=pflash,format=raw,read-only,file=$PREFIX/share/qemu/edk2-aarch64-code.fd
-netdev user,id=n1,hostfwd=tcp::2222-:22 -device virtio-net,netdev=n1
-cdrom alpine-virt-3.19.1-aarch64.iso
-nographic alpine.img

@raj-sharma-git I have a question.

Is there a special reason you want to run arm 64? It is still not accelerated - far slower then x86_64 ( I read about 2x or 4x slower - not sure about this). Do you "need" it?

I consider finding out how that works - I still don't understand the reason.

I thought that the performance would be better because it operates on the same architecture as the phone's processor. My assumption was that it would opt for virtualization instead of emulation, leading to superior performance.

Guys do you know about proot-distro and ubuntu distro? It's much lighter and faster than qemu and supports docker not sure about alpine

Edit: forget that, it has errors with cgroups and I didn't know how to fix it

I thought that the performance would be better because it operates on the same architecture as the phone's processor. My assumption was that it would opt for virtualization instead of emulation, leading to superior performance.

If you "need" arm 64 docker and can give me a good reason why, I might consider adding this to my repo. I really see no point at the moment. You might have. That's why I am asking you.

Guys do you know about proot-distro and ubuntu distro? It's much lighter and faster than qemu and supports docker not sure about alpine

Not true: https://en.wikipedia.org/wiki/Alpine_Linux

Alpine Linux is a Linux distribution designed to be small, simple, and secure. It uses musl, BusyBox, and OpenRC instead of the more commonly used glibc, GNU Core Utilities, and systemd.This makes Alpine one of few Linux distributions not to be based on the GNU Core Utilities.

I updated to alpine 3.20

  Error: Could not retrieve NBP file size from HTTP server.

  Error: Server response timeout.
BdsDxe: failed to load Boot0004 "UEFI HTTPv4 (MAC:525400123456)" from PciRoot(0x0)/Pci(0x2,0x0)/MAC(525400123456,0x1)/IPv4(0.0.0.0,0x0,DHCP,0.0.0.0,0.0.0.0,0.0.0.0)/Uri(): Not Found

what this?

_PXE-E16：未收到有效报价。BdsDxe ：无法从 PciRoot(0x0)/Pci(0x2,0x0)/MAC(525400123456,0x1)/IPv4(0.0.0.0,0x0,DHCP,0.0.0.0,0.0.0.0,0.0.0.0.0 加载 Boot0003“UEFI PXEv4 (MAC:525400123456)”：未找到
Did you solve it?

  Error: Could not retrieve NBP file size from HTTP server.

  Error: Server response timeout.
BdsDxe: failed to load Boot0004 "UEFI HTTPv4 (MAC:525400123456)" from PciRoot(0x0)/Pci(0x2,0x0)/MAC(525400123456,0x1)/IPv4(0.0.0.0,0x0,DHCP,0.0.0.0,0.0.0.0,0.0.0.0)/Uri(): Not Found

what this?

I think the Alpine version that is refered in this post is a bit outdated. Try using 3.19 or 3.20

这是来自QQ邮箱的假期自动回复邮件。你好，我最近正在休假中，无法亲自回复你的邮件。我将在假期结束后，尽快给你回复。

why install X86 VMs when phone‘s cup are now arm . Wouldn't installing an arm system run more efficiently?