Skip to content

anchorMC: Fail if required modulecmd is not found #2412

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
sawenzel merged 2 commits into from
Jul 16, 2026
+40 −1
Merged

anchorMC: Fail if required modulecmd is not found #2412

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6640bb2
anchorMC: Fail hard if modulecmd not found for 2-tag operations
sawenzel Jul 16, 2026
f11f17a
anchorMC: resolve modulecmd via readlink -f + /usr/share/Modules fall…
sawenzel Jul 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
41 changes: 40 additions & 1 deletion MC/run/ANCHOR/anchorMC.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,9 +178,48 @@ fi

export > env_base.env

# The 2-tag / alternative-reco setup below uses 'module save/purge/restore' to stash
# and swap the software environment. If the shell provides no 'module' function, build
# one on a 'modulecmd' binary. Its location varies by version, so resolve it flexibly
# (was hardwired to /usr/bin/modulecmd) and verify it supports saved collections. Fail
# hard -- but only when an alternative-reco tag is set, since single-tag jobs never call
# 'module' -- instead of silently no-op'ing every 'module' call (cf. O2-7070).
# The candidate list also resolves symlinks ('readlink -f') and probes the canonical
# environment-modules install (/usr/share/Modules): on some HPC/container setups
# /usr/bin/modulecmd is a *dangling* symlink (-> /etc/alternatives -> the real .tcl),
# which exec's as "No such file or directory"; this recovers when the real binary is
# still reachable, and otherwise still fails hard (a genuinely missing Modules tree in
# the job's mount namespace is a site-side problem).
if ! declare -F module > /dev/null; then
MODULECMD_BIN=""
for _cand in "${MODULECMD}" "/usr/bin/modulecmd" "$(command -v modulecmd 2>/dev/null)" \
"${MODULESHOME:+${MODULESHOME}/libexec/modulecmd}" \
"$(readlink -f /usr/bin/modulecmd 2>/dev/null)" \
/usr/share/Modules/libexec/modulecmd* /usr/share/Modules/*/libexec/modulecmd*; do
if [ -n "${_cand}" ] && [ -x "${_cand}" ] && "${_cand}" bash savelist > /dev/null 2>&1; then
MODULECMD_BIN="${_cand}"
break
fi
done
if [ -n "${MODULECMD_BIN}" ]; then
echo_info "Using modulecmd at ${MODULECMD_BIN}"
elif [ "${ALIEN_JDL_O2DPG_ASYNC_RECO_TAG}" ]; then
echo_error "Alternative-reco (2-tag) requested (ALIEN_JDL_O2DPG_ASYNC_RECO_TAG=${ALIEN_JDL_O2DPG_ASYNC_RECO_TAG})"
echo_error "but no 'modulecmd' supporting saved collections (save/restore) was found."
echo_error " Looked at: \$MODULECMD='${MODULECMD}', /usr/bin/modulecmd (+readlink -f), PATH,"
echo_error " \$MODULESHOME/libexec/modulecmd, /usr/share/Modules/{,*/}libexec/modulecmd*."
echo_error " Diagnostics: MODULESHOME='${MODULESHOME}' /usr/bin/modulecmd -> '$(readlink -f /usr/bin/modulecmd 2>/dev/null)' PATH='${PATH}'"
exit 1
else
echo_info "No 'modulecmd' found, but no alternative-reco tag is set, so 'module' is not needed; continuing."
fi
export MODULECMD_BIN
module() {
eval "$(/usr/bin/modulecmd bash "$@")";
if [ -z "${MODULECMD_BIN}" ]; then
echo_error "'module ${*}' invoked but no usable 'modulecmd' was found (see earlier diagnostics)."
return 127
fi
eval "$("${MODULECMD_BIN}" bash "$@")";
}
export -f module
fi
Expand Down
Loading