Skip to content

CollectiveX: name Slurm allocations after the GHA runner (dash correlation) #2762

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
Oseltamivir merged 3 commits into from
Aug 28, 2026
Merged

CollectiveX: name Slurm allocations after the GHA runner (dash correlation) #2762

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
e225d68
CollectiveX: name Slurm allocations after the GHA runner for dashboar…
Oseltamivir Aug 27, 2026
ca93f20
remove
Oseltamivir Aug 28, 2026
d27f05f
rem
Oseltamivir Aug 28, 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
14 changes: 14 additions & 0 deletions experimental/CollectiveX/runtime/common.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -424,10 +424,24 @@ exec python3 bench/run_ep.py "$@"
BASH
}

# Slurm job name for dashboard correlation. inferencex-dash's cluster collector
# samples allocation names (sacct JobName) and the dashboard joins them against
# the GHA job's runner_name — the production runners' convention
# (--job-name="$RUNNER_NAME"). Fall back to a fixed label for hand-driven runs;
# reject anything that is not a plain Slurm-safe token rather than quoting it.
collx_slurm_job_name() {
local name="${RUNNER_NAME:-}"
[[ "$name" =~ ^[A-Za-z0-9][A-Za-z0-9._-]{0,127}$ ]] || name="collectivex"
printf '%s' "$name"
}

# Allocate via salloc's stable grant message and assign JOB_ID in this shell.
# Record it so workflow cleanup can release a launcher interrupted by Actions.
# The job name is prepended so a caller-supplied --job-name still wins (salloc
# takes the last occurrence).
collx_salloc_jobid() {
local log_label=scheduler-allocation log job_id root="${COLLX_JOB_ROOT:-}"
set -- --job-name="$(collx_slurm_job_name)" "$@"
case "${COLLX_SALLOC_ATTEMPT:-1}" in
1) ;;
2|3) log_label+="-a${COLLX_SALLOC_ATTEMPT}" ;;
Expand Down