Differences

This shows you the differences between two versions of the page.

--- linux:docker [2023/11/14 00:07] – Wuff
+++ linux:docker [2025/09/04 00:12] (current) – [GPU hardware acceleration] Wuff
@@ Line 2: / Line 2: @@
 https://docs.docker.com/storage/volumes/
+===== Main install =====
+<code>
+apt-get -y install curl unzip dialog
+curl -fSSL get.docker.com | sh
+</code>
+Useful alias to exclude container overlays from the df command for use in .bashrc or .profile:
+<code>
+alias df='/usr/bin/df -x overlay'
+</code>
+===== Show logs =====
+<code>
+docker logs --follow <containername>
+</code>
 ===== Build image =====
@@ Line 103: / Line 121: @@
 docker tag current/image:tag new/image:tag
 </code>
+===== cmd shell of container =====
+<code>
+docker exec -it <mycontainer> bash
+docker exec -it <mycontainer> /bin/sh
+</code>
+===== GPU hardware acceleration =====
+Make sure the relevant GPU drivers are installed on the base system, then pass through the device via docker compose.
+Check devices are available, there should be a device per GPU starting at renderD128 for the first GPU:
+<code>
+ls -la /dev/dri
+</code>
+<code>
+  devices:
+   - /dev/dri:/dev/dri
+</code>
+for AMD:
+https://rocm.docs.amd.com/projects/install-on-linux/en/latest/install/install-methods/package-manager/package-manager-debian.html
+https://rocm.docs.amd.com/projects/install-on-linux/en/latest/how-to/docker.html
+<code>
+mkdir --parents --mode=0755 /etc/apt/keyrings
+wget https://repo.radeon.com/rocm/rocm.gpg.key -O - | gpg --dearmor | sudo tee /etc/apt/keyrings/rocm.gpg > /dev/null
+echo "deb [arch=amd64 signed-by=/etc/apt/keyrings/rocm.gpg] https://repo.radeon.com/rocm/apt/6.4.3 jammy main" > tee /etc/apt/sources.list.d/rocm.list
+echo -e 'Package: *\nPin: release o=repo.radeon.com\nPin-Priority: 600' > /etc/apt/preferences.d/rocm-pin-600
+apt update
+apt install rocm
+</code>
+Then add to the docker containers:
+<code>
+  devices:
+    - /dev/kfd
+    - /dev/dri
+  security_opt:
+    - seccomp=unconfined
+  environment:
+    - LIBVA_DRIVER_NAME=radeonsi
+</code>
+Alpine Linux docker images may require libva to be installed:
+<code>
+    post_start:
+      # for 13th gen Iris Xe GPUs use intel-media-driver for older gens libva-intel-driver
+      # - command: apk add libva-intel-driver intel-media-driver
+      #   user: root
+      # For AMD
+      - command: apk add libva libva-utils mesa-va-gallium
+        user: root
+</code>
+nvtop can be used for AMD and NVidia gpus to see if the gpu is being used.
+On Debian, nvtop is in the contrib repositories which need to be added.
+<code>
+#bookworm
+sed -r 's/^deb(.*)$/deb\1 contrib/g' /etc/apt/sources.list
+#new deb822 sources format
+sed -i 's/^Components: main$/& contrib non-free non-free-firmware/' /etc/apt/sources.list.d/debian.sources
+apt update
+</code>
+On Ubuntu systems, nvtop is in the default repositories.
+<code>
+apt-get install nvtop
+</code>
+===== docker ps short =====
+Short form of docker ps to only get the container names:
+<code>
+docker ps | sed 's/   */#/g' | cut -d "#" -f 7
+</code>
+===== docker command template =====
+https://stackoverflow.com/questions/50667371/docker-ps-output-formatting-list-only-names-of-running-containers
+===== docker running out of network addresses =====
+Error response from daemon: all predefined address pools have been fully subnetted
+<code json /etc/docker/daemon.json>
+{
+"default-address-pools":[
+    {"base":"169.254.2.0/23","size":28},
+    {"base":"169.254.4.0/22","size":28},
+    {"base":"169.254.8.0/21","size":28},
+    {"base":"169.254.16.0/20","size":28},
+    {"base":"169.254.32.0/19","size":28},
+    {"base":"169.254.64.0/18","size":28},
+    {"base":"169.254.128.0/18","size":28},
+    {"base":"169.254.192.0/19","size":28},
+    {"base":"169.254.224.0/20","size":28},
+    {"base":"169.254.240.0/21","size":28},
+    {"base":"169.254.248.0/22","size":28}
+  ]
+}
+</code>
+<code>
+docker restart
+docker network prune
+</code>