Skip to content

fix(runners): import B300 container images on the login host #2893

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

Open
cquil11 wants to merge 3 commits into
base: main
Choose a base branch
from
+2 −9
Open

fix(runners): import B300 container images on the login host #2893

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8d479d1
fix(runners): import B300 images on the login host [skip-sweep]
cquil11 Sep 8, 2026
9677323
test: remove B300 image import tests [skip-sweep]
cquil11 Sep 8, 2026
fffdaa1
fix(runners): simplify B300 import to a local call [skip-sweep]
cquil11 Sep 8, 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
11 changes: 2 additions & 9 deletions runners/launch_b300-dsxe.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -76,13 +76,7 @@ set -x
#
# Import a container image into the shared squash dir. Concurrent callers target the
# same path, so serialize on a per-file lock and skip when a valid squash file exists.
# --time bounds the step; an unbounded srun hangs the job if its step is lost.
#
# The import itself must run on a compute node: enroot builds the squashfs over an
# overlay mount, which the shared filesystem cannot back, and the login host is too
# small to unpack a multi-GB image. Reading the finished file is just I/O, so probe
# it here first -- a warm cache then costs no Slurm allocation at all. The in-srun
# check under the lock stays authoritative, so a stale probe only costs one step.
# Run the import locally on the runner/login host using its Enroot configuration.
import_squash_image() {
local image_ref="$1"
local sqsh="$2"
Expand All @@ -93,8 +87,7 @@ import_squash_image() {
return 0
fi

srun -N 1 -A "$SLURM_ACCOUNT" -p "$SLURM_PARTITION" \
--time="${ENROOT_IMPORT_TIME_LIMIT:-120}" bash -c "

@cursor cursor Bot Sep 8, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Import timeout no longer enforced

Medium Severity

Replacing srun --time with unbounded bash -c drops the import deadline. A stalled enroot import now holds the per-image lock indefinitely, and other runners wait on flock for an hour before failing while the hung import keeps the login-host runner busy.

Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit fffdaa1. Configure here.

@cursor cursor Bot Sep 8, 2026

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Login-host imports can exhaust resources

High Severity

Cold enroot import now runs on the 30 GiB login host with Enroot’s defaults (nproc processors, 10 download connections) and only a per-image lock. A production-size image, or two different images at once, can OOM or fill local disk and take down all 18 GHA runners on that host.

Fix in Cursor Fix in Web

Reviewed by Cursor Bugbot for commit fffdaa1. Configure here.

bash -c "
set -euo pipefail
exec 9>\"$lock\"
flock -w 3600 9
Expand Down