EPS
← All tasks·#1049Completed

daily-fix: document pod-to-HF upload wedge escalation ladder

kind: infra#wf-fix#wf-fix-fp:161ae961eea5#daily-auto-filed
track:

Highlight text on any card (body, plan, or an event) to anchor a comment, or leave a whole-task comment here. Mention @claude to summon a reply.

No comments yet.

Overview / Motivation

Auto-filed by /daily 2026-07-04 (route 2: behavior/logic change -> independent review), from the nightly transcript problem sweep.

Goal

Document the full three-rung wedge escalation ladder (HF_XET_DISABLE=1 -> HF_HUB_ENABLE_HF_TRANSFER=0 -> rsync-to-VM reroute) in upload-policy.md next to the #515 note so the next wedged session runs the ladder instead of rediscovering it.

Workflow gap

  • Bug observed: #931 spent ~50 min on 2026-07-04 breaking three successive pod-side HF upload wedges (xet stall -> hf_transfer stall -> CDN-route-specific dead end) before landing the working recovery: rsync 9.8 GB to the VM and upload from there. The #515 xet-CDN workaround note covers only the first rung.
  • Why it is a workflow gap: the failure originates in the workflow surface / helper named below, not in any one experiment.
  • Confidence (emitter): medium

Scope / surfaces

  • Primary target: .claude/rules/upload-policy.md
  • Session: 2f37cc81 (#931, ~06:00-06:52 UTC).

Constraints / invariants

  • Workflow-surface only; scripts/workflow_lint.py --check-asks + --check-references stay green; ruff clean on touched files.

Provenance

  • workflow_fix_target: .claude/rules/upload-policy.md
  • source: /daily 2026-07-04 problem sweep (transcript-mined)
Activity