ceph/container/Containerfile

ARG FROM_IMAGE="quay.io/centos/centos:stream9"
FROM $FROM_IMAGE

# allow FROM_IMAGE to be visible inside this stage
ARG FROM_IMAGE

# Ceph branch name
ARG CEPH_REF="main"

# Ceph SHA1
ARG CEPH_SHA1

# Ceph git repo (ceph-ci.git or ceph.git)
ARG CEPH_GIT_REPO

# (optional) Define the baseurl= for the ganesha.repo
ARG GANESHA_REPO_BASEURL="https://buildlogs.centos.org/centos/\$releasever-stream/storage/\$basearch/nfsganesha-9/"

# (optional) Set to "crimson" to install crimson packages.
ARG OSD_FLAVOR="default"

# (optional) Should be 'true' for CI builds (pull from shaman, etc.)
ARG CI_CONTAINER="true"

# (optional, for hackers) Use the following URL for ceph repo.
ARG CUSTOM_CEPH_REPO_URL

RUN /bin/echo -e "\
FROM_IMAGE: ${FROM_IMAGE}\n\
CEPH_REF: ${CEPH_REF}\n\
GANESHA_REPO_BASEURL: ${GANESHA_REPO_BASEURL} \n\
OSD_FLAVOR: ${OSD_FLAVOR} \n\
CI_CONTAINER: ${CI_CONTAINER}"

# Other labels are set automatically by container/build github action
# See: https://github.com/opencontainers/image-spec/blob/main/annotations.md
LABEL org.opencontainers.image.authors="Ceph Release Team <ceph-maintainers@ceph.io>" \
      org.opencontainers.image.documentation="https://docs.ceph.com/"

LABEL \
ceph=True \
FROM_IMAGE=${FROM_IMAGE} \
CEPH_REF=${CEPH_REF} \
CEPH_SHA1=${CEPH_SHA1} \
CEPH_GIT_REPO=${CEPH_GIT_REPO} \
GANESHA_REPO_BASEURL=${GANESHA_REPO_BASEURL} \
OSD_FLAVOR=${OSD_FLAVOR}

# Use bash so var setting works.
SHELL ["/bin/bash", "-c"]

# Derive EL_VER (el9/el10) and DIST_PATH (centos/9|rocky/10|almalinux/X)
# from /etc/os-release for later use in repo URLs.
RUN set -euo pipefail; \
    source /etc/os-release; \
    MAJOR="${VERSION_ID%%.*}"; \
    case "${ID}" in \
      centos)    DIST_PATH="centos/${MAJOR}" ;; \
      rocky)     DIST_PATH="rocky/${MAJOR}" ;; \
      almalinux) DIST_PATH="almalinux/${MAJOR}" ;; \
      *) echo "Unsupported base: ID=${ID} VERSION_ID=${VERSION_ID}" >&2; exit 1 ;; \
    esac; \
    EL_VER="el${MAJOR}"; \
    printf 'EL_VER=%s\nDIST_PATH=%s\nID=%s\nVERSION_ID=%s\nMAJOR=%s\n' \
           "$EL_VER" "$DIST_PATH" "$ID" "$VERSION_ID" "$MAJOR" > /etc/ceph-distro.env

#===================================================================================================
# Install ceph and dependencies, and clean up
# IMPORTANT: in official builds, use '--squash' build option to keep image as small as possible
#   keeping run steps separate makes local rebuilds quick, but images are big without squash option
#===================================================================================================

# Disable documentation
# (assumes only [main] section exists)
RUN \
    if grep -q 'tsflags' /etc/dnf/dnf.conf ; then \
        sed -i 's/tsflags=.*/tsflags=nodocs/g' /etc/dnf/dnf.conf ; \
    else \
        echo "tsflags=nodocs" >> /etc/dnf/dnf.conf ; \
    fi

# Pre-reqs
RUN dnf install -y --setopt=install_weak_deps=False epel-release jq

# NFS-Ganesha repo
RUN set -eux; \
    source /etc/ceph-distro.env; \
    if [[ "${DIST_PATH}" == rocky/* ]]; then \
        curl -fs -L "https://shaman.ceph.com/api/repos/nfs-ganesha/main/latest/${DIST_PATH}/repo?arch=$(arch)" -o /etc/yum.repos.d/ganesha.repo; \
    else \
      { \
        printf '%s\n' '[ganesha]'; \
        printf '%s\n' 'name=ganesha'; \
        printf '%s\n' "baseurl=${GANESHA_REPO_BASEURL}"; \
        printf '%s\n' 'gpgcheck=0'; \
        printf '%s\n' 'enabled=1'; \
      } > /etc/yum.repos.d/ganesha.repo; \
    fi

# ISCSI repo
RUN set -eux && \
    source /etc/ceph-distro.env && \
    curl -s -L https://shaman.ceph.com/api/repos/tcmu-runner/main/latest/${DIST_PATH}/repo?arch=$(arch) -o /etc/yum.repos.d/tcmu-runner.repo && \
    case "${CEPH_REF}" in \
        quincy|reef) \
            curl -fs -L https://download.ceph.com/ceph-iscsi/3/rpm/${EL_VER}/ceph-iscsi.repo -o /etc/yum.repos.d/ceph-iscsi.repo ;\
            ;;\
        main|*) \
            curl -fs -L https://shaman.ceph.com/api/repos/ceph-iscsi/main/latest/${DIST_PATH}/repo -o /etc/yum.repos.d/ceph-iscsi.repo ;\
            ;;\
    esac

# Ceph repo
RUN --mount=type=secret,id=prerelease_creds set -ex && \
    source /etc/ceph-distro.env && \
    if [ "$CUSTOM_CEPH_REPO_URL" ]; then \
        curl -L -o /tmp/custom-ceph.repo "$CUSTOM_CEPH_REPO_URL" && \
        mv /tmp/custom-ceph.repo  /etc/yum.repos.d/custom-ceph.repo && \
        exit 0 ; \
    fi && \
    rpm --import 'https://download.ceph.com/keys/release.asc' && \
    ARCH=$(arch); if [ "${ARCH}" == "aarch64" ]; then ARCH="arm64"; fi ;\
    IS_RELEASE=0 ;\
    if [[ "${CI_CONTAINER}" == "true" ]] ; then \
        # TODO: this can return different ceph builds (SHA1) for x86 vs. arm runs. is it important to fix?
        REPO_URL=$(curl -fs "https://shaman.ceph.com/api/search/?project=ceph&distros=${DIST_PATH}/${ARCH}&flavor=${OSD_FLAVOR}&ref=${CEPH_REF}&sha1=latest" | jq -r .[0].url) ;\
    else \
        IS_RELEASE=1 ;\
        source /run/secrets/prerelease_creds; \
        REPO_URL="https://${PRERELEASE_USERNAME}:${PRERELEASE_PASSWORD}@download.ceph.com/prerelease/ceph/rpm-${CEPH_REF}/${EL_VER}/" ;\
    fi && \
    rpm -Uvh "$REPO_URL/noarch/ceph-release-1-${IS_RELEASE}.${EL_VER}.noarch.rpm" ; \
    if [[ "$IS_RELEASE" == 1 ]] ; then \
        sed -i "s;http://download.ceph.com/;https://${PRERELEASE_USERNAME}:${PRERELEASE_PASSWORD}@download.ceph.com/prerelease/ceph/;" /etc/yum.repos.d/ceph.repo ; \
        dnf clean expire-cache ; \
    fi


# Copr repos
# scikit for mgr-diskprediction-local
# ref: https://github.com/ceph/ceph-container/pull/1821
RUN set -euo pipefail; \
    source /etc/ceph-distro.env; \
    if [ "$MAJOR" -le 9 ]; then \
        dnf install -y --setopt=install_weak_deps=False dnf-plugins-core && \
        dnf copr enable -y tchaikov/python-scikit-learn; \
    fi

# Update package mgr
RUN dnf update -y --setopt=install_weak_deps=False

# Define and install packages
# General
RUN echo "ca-certificates" > packages.txt
# Ceph
# TODO: remove lua-devel and luarocks once they are present in ceph.spec.in
#       ref: https://github.com/ceph/ceph/pull/54575#discussion_r1401199635
RUN echo \
"ceph-common \
ceph-exporter \
ceph-grafana-dashboards \
ceph-immutable-object-cache \
ceph-mds \
ceph-mgr-cephadm \
ceph-mgr-dashboard \
ceph-mgr-diskprediction-local \
ceph-mgr-k8sevents \
ceph-mgr-rook \
ceph-mgr \
ceph-mon \
ceph-osd \
ceph-radosgw lua-devel luarocks \
ceph-volume \
cephfs-mirror \
cephfs-top \
kmod \
libcephfs-daemon \
libradosstriper1 \
rbd-mirror" \
>> packages.txt

# Ceph "Recommends"
RUN set -euo pipefail; \
    echo "nvme-cli smartmontools" >> packages.txt; \
    source /etc/ceph-distro.env; \
    if [ "$MAJOR" -le 9 ]; then \
        echo "python3-saml" >> packages.txt; \
    fi

# NFS-Ganesha
RUN echo "\
dbus-daemon \
nfs-ganesha-ceph \
nfs-ganesha-rados-grace \
nfs-ganesha-rados-urls \
nfs-ganesha-rgw \
nfs-ganesha \
rpcbind \
sssd-client" >> packages.txt

# ISCSI
RUN echo "ceph-iscsi tcmu-runner python3-rtslib" >> packages.txt

# Ceph-CSI
# TODO: coordinate with @Madhu-1 to have Ceph-CSI install these itself if unused by ceph
#       @adk3798 does cephadm use these?
RUN echo "attr ceph-fuse rbd-nbd"  >> packages.txt

# Rook (only if packages must be in ceph container image)
RUN echo "systemd-udev" >> packages.txt

# Util packages (should be kept to only utils that are truly very useful)
# 'sgdisk' (from gdisk) is used in docs and scripts for clearing disks (could be a risk? @travisn @guits @ktdreyer ?)
# 'ps' (from procps-ng) and 'hostname' are very valuable for debugging and CI
# TODO: remove sg3_utils once they are moved to ceph.spec.in with libstoragemgmt
#       ref: https://github.com/ceph/ceph-container/pull/2013#issuecomment-1248606472
RUN echo "gdisk hostname procps-ng sg3_utils e2fsprogs lvm2 gcc" >> packages.txt

# scikit
RUN echo "python3-scikit-learn" >> packages.txt

# ceph-node-proxy
RUN echo "ceph-node-proxy" >> packages.txt

RUN echo "=== PACKAGES TO BE INSTALLED ==="; cat packages.txt
RUN echo "=== INSTALLING ===" ; \
dnf install -y --setopt=install_weak_deps=False --setopt=skip_missing_names_on_install=False --enablerepo=crb $(cat packages.txt)

# XXX why isn't this done in the ganesha package?
RUN mkdir -p /var/run/ganesha

# Disable sync with udev since the container can not contact udev
RUN \
    sed -i -e 's/udev_rules = 1/udev_rules = 0/' \
           -e 's/udev_sync = 1/udev_sync = 0/' \
           -e 's/obtain_device_list_from_udev = 1/obtain_device_list_from_udev = 0/' \
        /etc/lvm/lvm.conf && \
    # validate the sed command worked as expected
    grep -sqo "udev_sync = 0" /etc/lvm/lvm.conf && \
    grep -sqo "udev_rules = 0" /etc/lvm/lvm.conf && \
    grep -sqo "obtain_device_list_from_udev = 0" /etc/lvm/lvm.conf

# CLEAN UP!
RUN set -ex && \
    dnf clean all && \
    rm -rf /var/cache/dnf/* && \
    rm -rf /var/lib/dnf/* && \
    rm -f /var/lib/rpm/__db* && \
    # remove unnecessary files with big impact
    rm -rf /etc/selinux /usr/share/selinux && \
    rm -f /etc/yum.repos.d/{ceph,ganesha,tcmu-runner,ceph-iscsi}.repo

# Verify that the packages installed haven't been accidentally cleaned, then
# clean the package list and re-clean unnecessary RPM database files
RUN rpm -q $(cat packages.txt) && rm -f /var/lib/rpm/__db* && rm -f *packages.txt

#
# Set some envs in the container for quickly inspecting details about the build at runtime
ENV CEPH_IS_DEVEL="${CI_CONTAINER}" \
    CEPH_REF="${CEPH_REF}" \
    CEPH_VERSION="${CEPH_REF}" \
    CEPH_OSD_FLAVOR="${OSD_FLAVOR}" \
    FROM_IMAGE="${FROM_IMAGE}"