pull down to refresh

101 sats \ 1 reply \ @kepford OP 23 Jul \ parent \ on: bluefin - Next gen Linux workstation, reliability, performance, & sustainability devs
If an update goes bad, you can just boot into the previous image and you are back to where you started from
This to me is the biggest selling point for me and the workstation usecase. I'll have to take a look at bootc.
write
preview
265 sats \ 0 replies \ @freetx 23 Jul
Yes, there are two main benefits: (a) atomic updates that can be rolled back, and (b) Both your OS and the apps / containers that run on the OS uses the same tooling (ie. podman).
Most of the servers I roll-out now are bootc versions and all applications I container-ize. That means that composing both the OS and the apps on the servers all uses same workflow.
Example (toy example) of a simple server:
FROM quay.io/fedora/fedora-bootc:latest

# Install basic server packages
RUN dnf install -y \
    cockpit \
    firewalld \
    openssh-server \
    bash-completion \
    git \
    sysstat \
    wget \
    && dnf clean all

RUN useradd -m -G wheel core && \
    echo 'core:password123' | chpasswd

RUN systemctl enable sshd firewalld cockpit.socket

CMD ["/sbin/init"]

LABEL containers.bootc=1 
LABEL ostree.bootable=1 
LABEL bootc.build.iso=1
Then you can build it with something like:
sudo podman build -t "myserver" .
Then you can create your qcow2 image for use in qemu-kvm or even build an installable iso with bootc-image-builder, like:
sudo podman run --rm -it --privileged -v $(pwd)/output:/output \
  quay.io/centos-bootc/bootc-image-builder:latest \
  --type qcow2 \
  myserver

sudo podman run --rm -it --privileged -v $(pwd)/output:/output \
  quay.io/centos-bootc/bootc-image-builder:latest \
  --type iso \
  myserver
reply