Merge branch '2023.06-software.eessi.io' into 2023.06-software.eessi.…

…io_Paraver

.github/workflows/scorecards.yml

@@ -4,24 +4,20 @@
 
 name: Scorecards supply-chain security
 on:
-  # For Branch-Protection check. Only the default branch is supported. See
-  # https://github.com/ossf/scorecard/blob/main/docs/checks.md#branch-protection
-  branch_protection_rule:
   # To guarantee Maintained check is occasionally updated. See
   # https://github.com/ossf/scorecard/blob/main/docs/checks.md#maintained
   schedule:
     - cron: '25 15 * * 3'
   push:
-    branches: [ "main" ]
-  pull_request:
-    branches:    
-      - main
+    branches:
+      - '*-software.eessi.io'
 
 # Declare default permissions as read only.
 permissions: read-all
 
 jobs:
   analysis:
+    if: github.repository_owner == 'EESSI'   # Prevent running on forks
     name: Scorecards analysis
     runs-on: ubuntu-latest
     permissions:
@@ -35,12 +31,12 @@ jobs:
 
     steps:
       - name: "Checkout code"
-        uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
+        uses: actions/checkout@a5ac7e51b41094c92402da3b24376905380afc29 # v4.1.6
         with:
           persist-credentials: false
 
       - name: "Run analysis"
-        uses: ossf/scorecard-action@0864cf19026789058feabb7e87baa5f140aac736 # v2.3.1
+        uses: ossf/scorecard-action@dc50aa9510b46c811795eb24b2f1ba02a914e534 # v2.3.3
         with:
           results_file: results.sarif
           results_format: sarif
@@ -62,14 +58,14 @@ jobs:
       # Upload the results as artifacts (optional). Commenting out will disable uploads of run results in SARIF
       # format to the repository Actions tab.
       - name: "Upload artifact"
-        uses: actions/upload-artifact@3cea5372237819ed00197afe530f5a7ea3e805c8 # v3.1.0
+        uses: actions/upload-artifact@65462800fd760344b1a7b4382951275a0abb4808 # v4.3.3
         with:
           name: SARIF file
           path: results.sarif
           retention-days: 5
 
       # Upload the results to GitHub's code scanning dashboard.
       - name: "Upload to code-scanning"
-        uses: github/codeql-action/upload-sarif@807578363a7869ca324a79039e6db9c843e0e100 # v2.1.27
+        uses: github/codeql-action/upload-sarif@9fdb3e49720b44c48891d036bb502feb25684276 # v3.25.6
         with:
           sarif_file: results.sarif

.github/workflows/test-software.eessi.io.yml

@@ -1,11 +1,15 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Check for missing software installations in software.eessi.io
-on: [push, pull_request, workflow_dispatch]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
+  workflow_dispatch:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   check_missing:
-    runs-on: ubuntu-22.04
+    runs-on: ubuntu-latest
     strategy:
       fail-fast: false
       matrix:

.github/workflows/test_eessi_container_script.yml

@@ -1,11 +1,15 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Tests for eessi_container.sh script
-on: [push, pull_request, workflow_dispatch]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
+  workflow_dispatch:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   eessi_container_script:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     strategy:
       fail-fast: false
       matrix:
@@ -45,7 +49,8 @@ jobs:
               elif [[ ${{matrix.SCRIPT_TEST}} == 'listrepos_default' ]]; then
                 outfile=out_listrepos.txt
                 ./eessi_container.sh --verbose --list-repos | tee ${outfile}
-                grep "EESSI" ${outfile}
+                # make sure that the default EESSI software repository is available
+                grep "software.eessi.io" ${outfile}
 
               # test use of --list-repos with custom repos.cfg
               elif [[ ${{matrix.SCRIPT_TEST}} == 'listrepos_custom' ]]; then
@@ -57,11 +62,12 @@ jobs:
                 echo "[EESSI/20HT.TP]" >> cfg/repos.cfg
                 echo "repo_version = 20HT.TP" >> cfg/repos.cfg
                 ./eessi_container.sh --verbose --list-repos | tee ${outfile}
-                grep "EESSI" ${outfile}
+                # make sure that the default EESSI software repository is available
+                grep "software.eessi.io" ${outfile}
 
                 export EESSI_REPOS_CFG_DIR_OVERRIDE=${PWD}/cfg
                 ./eessi_container.sh --verbose --list-repos | tee ${outfile2}
-                grep "[EESSI/2023.02]" ${outfile2}
+                grep "EESSI/20AB.CD" ${outfile2}
 
               # test use of --mode run
               elif [[ ${{matrix.SCRIPT_TEST}} == 'run' ]]; then

.github/workflows/test_licenses.yml

@@ -1,11 +1,14 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Test software licenses
-on: [push, pull_request]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     steps:
         - name: Check out software-layer repository
           uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1

.github/workflows/tests.yml

@@ -1,14 +1,17 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Tests
-on: [push, pull_request]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     strategy:
       matrix:
-        python: [3.6, 3.7, 3.8, 3.9, '3.10']
+        python: [3.7, 3.8, 3.9, '3.10']
       fail-fast: false
     steps:
     - name: checkout

.github/workflows/tests_archdetect.yml

@@ -1,11 +1,14 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Tests for eessi_archdetect.sh
-on: [push, pull_request]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     strategy:
       matrix:
         proc_cpuinfo:

.github/workflows/tests_init.yml

@@ -1,14 +1,17 @@
 # documentation: https://help.github.com/en/articles/workflow-syntax-for-github-actions
 name: Tests for init scripts
-on: [push, pull_request]
+on:
+  push:
+    branches: [ "*-software.eessi.io" ]
+  pull_request:
 permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     strategy:
       matrix:
-        python: [3.6, 3.7, 3.8, 3.9, '3.10']
+        python: [3.7, 3.8, 3.9, '3.10']
       fail-fast: false
     steps:
     - name: checkout

.github/workflows/tests_readme.yml

@@ -14,7 +14,7 @@ permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     steps:
         - name: Check out software-layer repository
           uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1

.github/workflows/tests_scripts.yml

@@ -28,7 +28,7 @@ permissions:
   contents: read # to fetch code (actions/checkout)
 jobs:
   build:
-    runs-on: ubuntu-20.04
+    runs-on: ubuntu-latest
     steps:
     - name: checkout
       uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1

EESSI-extend-2023.06-easybuild.eb

@@ -139,12 +139,17 @@ setenv ("EASYBUILD_SYSROOT", sysroot)
 setenv ("EASYBUILD_PREFIX", pathJoin(working_dir, "easybuild"))
 setenv ("EASYBUILD_INSTALLPATH", easybuild_installpath)
 setenv ("EASYBUILD_HOOKS", pathJoin(os.getenv("EESSI_PREFIX"), "init", "easybuild", "eb_hooks.py"))
-setenv ("EASYBUILD_UMASK", "002")
+
+-- Make sure to use the general umask that allows a global read
+setenv ("EASYBUILD_UMASK", "022")
+
+-- Allow this module to be loaded when running EasyBuild
+setenv ("EASYBUILD_ALLOW_LOADED_MODULES", "EasyBuild,EESSI-extend")
 
 -- Set all related environment variables if we have project or user installations (including extending MODULEPATH)
 if (user_modulepath ~= nil) then
   -- Use a more restrictive umask for this case
-  setenv ("EASYBUILD_UMASK", "022")
+  setenv ("EASYBUILD_UMASK", "077")
   setenv ("EASYBUILD_STICKY_BIT", "1")
   -- configure MODULEPATH
   if (project_modulepath ~= nil) then
@@ -155,6 +160,7 @@ elseif (project_modulepath ~= nil) then
   setenv ("EASYBUILD_SET_GID_BIT", "1")
   setenv ("EASYBUILD_GROUP_WRITABLE_INSTALLDIR", "1")
   setenv ("EASYBUILD_STICKY_BIT", "0")
+  setenv ("EASYBUILD_UMASK", "002")
   -- configure MODULEPATH
   prepend_path("MODULEPATH", project_modulepath)
 end

EESSI-install-software.sh

@@ -17,6 +17,11 @@ display_help() {
   echo "  --skip-cuda-install    -  disable installing a full CUDA SDK in the host_injections prefix (e.g. in CI)"
 }
 
+# Function to check if a command exists
+function command_exists() {
+    command -v "$1" >/dev/null 2>&1
+}
+
 function copy_build_log() {
     # copy specified build log to specified directory, with some context added
     build_log=${1}
@@ -238,10 +243,11 @@ else
     echo "Skipping installation of CUDA SDK in host_injections, since the --skip-cuda-install flag was passed OR no EasyBuild module was found"
 fi
 
-# Install drivers in host_injections
-# TODO: this is commented out for now, because the script assumes that nvidia-smi is available and works;
-#       if not, an error is produced, and the bot flags the whole build as failed (even when not installing GPU software)
-# ${EESSI_PREFIX}/scripts/gpu_support/nvidia/link_nvidia_host_libraries.sh
+# Install NVIDIA drivers in host_injections (if they exist)
+if command_exists "nvidia-smi"; then
+    echo "Command 'nvidia-smi' found. Installing NVIDIA drivers for use in prefix shell..."
+    ${EESSI_PREFIX}/scripts/gpu_support/nvidia/link_nvidia_host_libraries.sh
+fi
 
 # use PR patch file to determine in which easystack files stuff was added
 changed_easystacks=$(cat ${pr_diff} | grep '^+++' | cut -f2 -d' ' | sed 's@^[a-z]/@@g' | grep '^easystacks/.*yml$' | egrep -v 'known-issues|missing') 

bot/build.sh

@@ -138,7 +138,7 @@ echo "bot/build.sh: EESSI_VERSION_OVERRIDE='${EESSI_VERSION_OVERRIDE}'"
 # determine CVMFS repo to be used from .repository.repo_name in ${JOB_CFG_FILE}
 # here, just set EESSI_CVMFS_REPO_OVERRIDE, a bit further down
 # "source init/eessi_defaults" via sourcing init/minimal_eessi_env
-export EESSI_CVMFS_REPO_OVERRIDE=$(cfg_get_value "repository" "repo_name")
+export EESSI_CVMFS_REPO_OVERRIDE=/cvmfs/$(cfg_get_value "repository" "repo_name")
 echo "bot/build.sh: EESSI_CVMFS_REPO_OVERRIDE='${EESSI_CVMFS_REPO_OVERRIDE}'"
 
 # determine architecture to be used from entry .architecture in ${JOB_CFG_FILE}
@@ -223,7 +223,15 @@ mkdir -p ${TARBALL_TMP_BUILD_STEP_DIR}
 BUILD_STEP_ARGS+=("--save" "${TARBALL_TMP_BUILD_STEP_DIR}")
 BUILD_STEP_ARGS+=("--storage" "${STORAGE}")
 # add options required to handle NVIDIA support
-BUILD_STEP_ARGS+=("--nvidia" "all")
+if command_exists "nvidia-smi"; then
+    echo "Command 'nvidia-smi' found, using available GPU"
+    BUILD_STEP_ARGS+=("--nvidia" "all")
+else
+    echo "No 'nvidia-smi' found, no available GPU but allowing overriding this check"
+    BUILD_STEP_ARGS+=("--nvidia" "install")
+fi
+# Retain location for host injections so we don't reinstall CUDA
+# (Always need to run the driver installation as available driver may change)
 if [[ ! -z ${SHARED_FS_PATH} ]]; then
     BUILD_STEP_ARGS+=("--host-injections" "${SHARED_FS_PATH}/host-injections")
 fi

bot/check-test.sh

@@ -101,7 +101,7 @@ elif [[ ${ERROR} -eq 1 ]]; then
     reason="EESSI test suite was not run, test step itself failed to execute."
     status="FAILURE"
 else
-    summary=":grin: FAILURE"
+    summary=":cry: FAILURE"
     reason="Failed for unknown reason"
     status="FAILURE"
 fi

bot/inspect.sh

@@ -27,6 +27,10 @@
 # stop as soon as something fails
 set -e
 
+# script_dir is the directory that contains THIS (inspect.sh) script, usually
+# stored in the directory '.../bot'
+script_dir=$(dirname $(realpath $BASH_SOURCE))
+
 display_help() {
   echo "usage: $0 [OPTIONS]"
   echo "  -h | --help            -  display this usage information"
@@ -81,8 +85,8 @@ done
 set -- "${POSITIONAL_ARGS[@]}"
 
 # source utils.sh and cfg_files.sh
-source scripts/utils.sh
-source scripts/cfg_files.sh
+source ${script_dir}/../scripts/utils.sh
+source ${script_dir}/../scripts/cfg_files.sh
 
 if [[ -z ${resume_tgz} ]]; then
     echo_red "path to tarball for resuming build job is missing"
@@ -255,10 +259,8 @@ CMDLINE_ARGS+=("--storage" "${JOB_STORAGE}")
 
 # make sure some environment settings are available inside the shell started via
 # startprefix
-base_dir=$(dirname $(realpath $0))
-# base_dir of inspect.sh script is '.../bot', 'init' dir is at the same level
 # TODO better use script from tarball???
-source ${base_dir}/../init/eessi_defaults
+source ${script_dir}/../init/eessi_defaults
 
 if [ -z $EESSI_VERSION ]; then
     echo "ERROR: \$EESSI_VERSION must be set!" >&2
@@ -432,14 +434,14 @@ echo "Executing command to start interactive session to inspect build job:"
 # These initializations are combined into a single script that is executed when
 # the shell in startprefix is started. We set the env variable BASH_ENV here.
 if [[ -z ${run_in_prefix} ]]; then
-    echo "./eessi_container.sh ${CMDLINE_ARGS[@]}"
+    echo "${script_dir}/../eessi_container.sh ${CMDLINE_ARGS[@]}"
     echo "                     -- ${EESSI_COMPAT_LAYER_DIR}/startprefix"
-    ./eessi_container.sh "${CMDLINE_ARGS[@]}" \
+    ${script_dir}/../eessi_container.sh "${CMDLINE_ARGS[@]}" \
                      -- ${EESSI_COMPAT_LAYER_DIR}/startprefix
 else
-    echo "./eessi_container.sh ${CMDLINE_ARGS[@]}"
+    echo "${script_dir}/../eessi_container.sh ${CMDLINE_ARGS[@]}"
     echo "                     -- ${EESSI_COMPAT_LAYER_DIR}/startprefix <<< ${run_in_prefix}"
-    ./eessi_container.sh "${CMDLINE_ARGS[@]}" \
+    ${script_dir}/../eessi_container.sh "${CMDLINE_ARGS[@]}" \
                      -- ${EESSI_COMPAT_LAYER_DIR}/startprefix <<< ${run_in_prefix}
 fi
 

bot/test.sh

@@ -160,7 +160,7 @@ echo "bot/test.sh: EESSI_PILOT_VERSION_OVERRIDE='${EESSI_PILOT_VERSION_OVERRIDE}
 # determine CVMFS repo to be used from .repository.repo_name in ${JOB_CFG_FILE}
 # here, just set EESSI_CVMFS_REPO_OVERRIDE, a bit further down
 # "source init/eessi_defaults" via sourcing init/minimal_eessi_env
-export EESSI_CVMFS_REPO_OVERRIDE=$(cfg_get_value "repository" "repo_name")
+export EESSI_CVMFS_REPO_OVERRIDE=/cvmfs/$(cfg_get_value "repository" "repo_name")
 echo "bot/test.sh: EESSI_CVMFS_REPO_OVERRIDE='${EESSI_CVMFS_REPO_OVERRIDE}'"
 
 # determine architecture to be used from entry .architecture in ${JOB_CFG_FILE}
@@ -204,6 +204,9 @@ if [[ -z ${RESUME_DIR} ]]; then
 else
   TEST_STEP_ARGS+=("--resume" "${RESUME_DIR}")
 fi
+# Bind mount /sys/fs/cgroup so that we can determine the amount of memory available in our cgroup for
+# Reframe configuration
+TEST_STEP_ARGS+=("--extra-bind-paths" "/sys/fs/cgroup:/hostsys/fs/cgroup:ro")
 
 # prepare arguments to test_suite.sh (specific to test step)
 declare -a TEST_SUITE_ARGS=()

configure_easybuild

@@ -32,6 +32,12 @@ if [[ "$EESSI_CPU_FAMILY" == "aarch64" ]]; then
     DEPS_TO_FILTER="${DEPS_TO_FILTER},Yasm"
 fi
 
+# Version 23.06 of EESSI ships PSM2 in the compat layer, so we can filter this out while retaining support for OFA fabric
+# (longer term this is probably not the right move as PSM2 should be configured with accelerator support, hence the restricted version)
+if [[ "$EESSI_VERSION" == "2023.06" ]]; then
+    DEPS_TO_FILTER="${DEPS_TO_FILTER},PSM2"
+fi
+
 export EASYBUILD_FILTER_DEPS=$DEPS_TO_FILTER
 
 export EASYBUILD_MODULE_EXTENSIONS=1