Skip to content

fix(preflight): score an empty node set as an error - #694

Merged
hnimra-amd merged 1 commit into
release/cvs-0.3.0from
hnimrama/fix-AIMVT-436-0.3.0-pick
Oct 30, 2026
Merged

hnimra-amd merged 1 commit into
release/cvs-0.3.0from
hnimrama/fix-AIMVT-436-0.3.0-pick

Conversation

@hnimra-amd

Copy link
Copy Markdown
Collaborator

Summary

  • Cherry-pick of the squash merge of fix(preflight): score an empty node set as an error #679 onto release/cvs-0.3.0.
  • Preflight summarisers treat status: SKIPPED as skipped even when the skipped flag is absent, so a crash backfill is no longer scored as PASS 0/0.
  • A check that ran and tested no nodes is reported as ERROR, and that error fails the preflight report.
  • IFoE L2, Node Smoke Tier 1, Node Smoke Tier 3, and SSH reachability use the same skip, blocked, and empty rules.
  • On this release branch, an ERROR check is logged at error level with FAIL and BLOCKED.

https://amd-hub.atlassian.net/browse/AIMVT-436

Test

Checked the report summarizer on release/cvs-0.3.0 (6a43049) and on the same fix (c8de366, now this cherry-pick of 8a38dde8) with an empty node set.

On release/cvs-0.3.0 the empty result is scored as a pass, and the report stays PASS:

  • SSH reachability: PASS, 0/0 nodes reachable
  • Node Smoke Tier 1: PASS, 0/0
  • Node Smoke Tier 3: PASS, 0/0
  • IFoE L2: PASS, 0/0
  • A status-only SKIPPED backfill for IFoE, Tier 1, and Tier 3 is also PASS 0/0

On this fix those same inputs are ERROR (tested no nodes), and a status-only SKIPPED backfill stays SKIPPED. The empty-node unit tests pass (8/8). An empty SSH reachability result fails the preflight report.

A live preflight_checks run on two reachable nodes does not hit this path. Both release/cvs-0.3.0 and the fix finished PASS, 72 passed and 8 skipped, with SSH and node smoke at 2/2. Both logs also contain Node Smoke Tier 3 PASS on 0/2 nodes after a Tier 3 torchrun abort; the summary still records Tier 3 as 2/2. That line is unchanged by this fix.

A crash backfill records status SKIPPED without the skipped flag, and summarisers that ignored that status scored zero nodes as a pass.
@github-actions

Copy link
Copy Markdown
Security scan found no critical or high findings on lines changed by this pull request

1 medium, low, or unrated finding(s) are not listed; the scan run has them.

View the security scan run for complete reports.

“New” means the finding is located in a changed file on an added or modified line; it is not a base-versus-head findings delta.

@hnimra-amd
hnimra-amd merged commit 1e8e34c into release/cvs-0.3.0 Oct 30, 2026
14 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants