Upgrade container to OpenHPC 4.x/Rocky 10

Signed-off-by: Timothy Middelkoop <[email protected]>

Author: middelkoopt

.github/workflows/build-container.yml

@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - 3.x
+      - 4.x
     tags:
       - 'v*'
   workflow_dispatch:

containers/README.md

@@ -1,7 +1,7 @@
-# OpenHPC 3.x with Slurm running Rocky9 in a container
+# OpenHPC 4.x with Slurm running Rocky10 in a container
 
 This is a simple single-user container environment for learning and testing
-Slurm on OpenHPC 3.x with Rocky9.
+Slurm on OpenHPC 4.x with Rocky10.
 
 The cluster contains a head node, login node, and 8 compute nodes as separate
 containers with a shared docker network and shared docker storage.  The

containers/build.sh

@@ -10,7 +10,7 @@ echo "=== setup ${container}"
 
 set -e
 echo '=== build openhpc'
-"${container}" build -t openhpc/openhpc:3 -f openhpc/Containerfile openhpc \
+"${container}" build -t openhpc/openhpc:4 -f openhpc/Containerfile openhpc \
 	--build-arg USER="${user}" \
 	--build-arg ARCH="${arch/arm64/aarch64}"
 

containers/container/Containerfile

@@ -1,4 +1,4 @@
-FROM registry.docker.com/rockylinux/rockylinux:9
+FROM registry.docker.com/rockylinux/rockylinux:10
 ## This container is based on ../openhpc/Containerfile and ../node/Containerfile.
 ## The keeps it in a single file for simplicity, please keep containers in sync.
 ## The sections (delimited with a comment) are identical.  The scheduler has been removed.
@@ -7,12 +7,12 @@ FROM registry.docker.com/rockylinux/rockylinux:9
 
 ## Base install
 ARG ARCH=aarch64
-RUN dnf install -y http://repos.openhpc.community/OpenHPC/3/EL_9/$ARCH/ohpc-release-3-1.el9.$ARCH.rpm
+RUN dnf install -y http://repos.openhpc.community/OpenHPC/4/EL_10/$ARCH/ohpc-release-4-1.el10.$ARCH.rpm
 
 RUN dnf install -y dnf-plugins-core && \
     dnf config-manager --set-enabled crb && \
     dnf upgrade -y && \
-    dnf install -y unzip jq procps-ng iproute bind-utils findutils && \
+    dnf install -y unzip jq yq procps-ng iproute bind-utils findutils && \
     dnf install -y ohpc-base
 
 ## Common packages
@@ -31,12 +31,12 @@ RUN dnf install -y ohpc-base-compute lmod-ohpc \
     git patch file zstd bzip2 xz \
     python3 python3-pip \
     gcc-c++ gcc-gfortran \
-    ohpc-gnu14-runtimes gnu14-compilers-ohpc \
-    openmpi5-pmix-gnu14-ohpc openblas-gnu14-ohpc netcdf-gnu14-openmpi5-ohpc
+    ohpc-gnu15-runtimes gnu15-compilers-ohpc \
+    openmpi5-pmix-gnu15-ohpc openblas-gnu15-ohpc netcdf-gnu15-openmpi5-ohpc
 
 ## Compilers, runtimes, and libraries.
-RUN dnf install -y spack-ohpc ohpc-gnu14-openmpi5-parallel-libs \
-    ohpc-gnu14-python-libs ohpc-gnu14-runtimes
+RUN dnf install -y spack-ohpc ohpc-gnu15-openmpi5-parallel-libs \
+    ohpc-gnu15-python-libs ohpc-gnu15-runtimes
 
 ### Container configuration (unique to this container)
 

containers/examples/mpi/run.sh

@@ -1,5 +1,6 @@
 #!/bin/bash
+set -e
 
-module load gnu13 openmpi5 prun
+module load gnu15 openmpi5 prun
 mpicc mpi.c -o mpi
 sbatch sbatch.sh

containers/examples/mpi/sbatch.sh

@@ -1,8 +1,9 @@
 #!/bin/bash
-
 #SBATCH -n 8
 #SBATCH --tasks-per-node=1
 
-module load gnu13 openmpi5 prun
+set -e
+
+module load gnu15 openmpi5 prun
 
 prun ./mpi

containers/head/Containerfile

@@ -1,4 +1,4 @@
-FROM openhpc/openhpc:3
+FROM openhpc/openhpc:4
 
 RUN dnf install -y ohpc-slurm-server
 

containers/node/Containerfile

@@ -1,20 +1,20 @@
-FROM openhpc/openhpc:3
+FROM openhpc/openhpc:4
 ## Update ../container/Containerfile with any changes made here.
 
 ## Base compute node
 RUN dnf install -y ohpc-base-compute lmod-ohpc \
     git patch file zstd bzip2 xz \
     python3 python3-pip \
     gcc-c++ gcc-gfortran \
-    ohpc-gnu14-runtimes gnu14-compilers-ohpc \
-    openmpi5-pmix-gnu14-ohpc openblas-gnu14-ohpc netcdf-gnu14-openmpi5-ohpc
+    ohpc-gnu15-runtimes gnu15-compilers-ohpc \
+    openmpi5-pmix-gnu15-ohpc openblas-gnu15-ohpc netcdf-gnu15-openmpi5-ohpc
 
 ## Scheduler
 RUN dnf install -y ohpc-slurm-client
 
 ## Compilers, runtimes, and libraries.
-RUN dnf install -y spack-ohpc ohpc-gnu14-openmpi5-parallel-libs \
-    ohpc-gnu14-python-libs ohpc-gnu14-runtimes
+RUN dnf install -y spack-ohpc ohpc-gnu15-openmpi5-parallel-libs \
+    ohpc-gnu15-python-libs ohpc-gnu15-runtimes
 
 RUN dnf clean all && \
     rm -rf /var/cache/dnf /var/log/dnf.librepo.log /var/log/dnf.rpm.log /var/log/hawkey.log

containers/openhpc/Containerfile

@@ -1,23 +1,25 @@
-FROM registry.docker.com/rockylinux/rockylinux:9
+FROM registry.docker.com/rockylinux/rockylinux:10
 ## Update ../container/Containerfile with any changes made here.
 
 ## Base install
 ARG ARCH=aarch64
-RUN dnf install -y http://repos.openhpc.community/OpenHPC/3/EL_9/$ARCH/ohpc-release-3-1.el9.$ARCH.rpm
+RUN dnf install -y http://repos.openhpc.community/OpenHPC/4/EL_10/$ARCH/ohpc-release-4-1.el10.$ARCH.rpm
 
 RUN dnf install -y dnf-plugins-core && \
     dnf config-manager --set-enabled crb && \
     dnf upgrade -y && \
-    dnf install -y unzip jq procps-ng iproute bind-utils findutils && \
+    dnf install -y unzip jq yq procps-ng iproute bind-utils findutils && \
     dnf install -y ohpc-base
 
-## Configure common image
-RUN dnf install -y munge && \
-    /usr/sbin/create-munge-key
-
 ## Common packages
 RUN dnf install -y iputils vim rsync unzip git
 
+## Scheduler
+RUN dnf install -y munge && \
+    /usr/sbin/mungekey && \
+    chown -v munge:munge /etc/munge/munge.key && \
+    install -v -dp --owner munge --group munge /run/munge
+
 ## Create user
 ARG USER=auser
 RUN useradd --create-home --shell=/bin/bash --user-group $USER && \

containers/run.sh

@@ -1,4 +1,5 @@
 #!/bin/bash
+set -e
 
 container=${CONTAINER:-docker}
 CONTAINER="${container}" ./build.sh
@@ -12,6 +13,7 @@ for I in {0..7}; do
 	"${container}" run -d --rm "${network}" "${volume}" --name=openhpc-node-"${I}" --hostname=node-"${I}" openhpc/node
 done
 
+set +e
 echo '=== Login Node (exit to shutdown cluster)'
 ./ssh.sh