Skip to content

kdump: avoid false negatives from fixed 100s stall window#4571

Merged
LiliDeng merged 1 commit into
mainfrom
baihua/kdump-adaptive-timeout
Jun 30, 2026
Merged

kdump: avoid false negatives from fixed 100s stall window#4571
LiliDeng merged 1 commit into
mainfrom
baihua/kdump-adaptive-timeout

Conversation

@lubaihua33

@lubaihua33 lubaihua33 commented Jun 30, 2026

Copy link
Copy Markdown
Collaborator

What
This PR replaces the fixed ~100s stall threshold in kdump result checking with adaptive windows derived from imeout_of_dump_crash.

Why
On large-memory VMs, vmcore write can pause/grow slowly and exceed a fixed 100s no-growth check, causing false failures (e.g., vmcore-incomplete).

Changes
Replace fixed max_tries = 20 with dynamic windows:

  • no-file window: max(100, min(timeout/4, 300))
  • no-growth window: max(100, min(timeout/3, 600))
  • Convert windows into retry counts with 5s polling interval.
  • Include effective wait duration in failure messages for diagnostics.

@lubaihua33 lubaihua33 requested a review from LiliDeng as a code owner June 30, 2026 08:51
Copilot AI review requested due to automatic review settings June 30, 2026 08:51
@github-actions

Copy link
Copy Markdown

✅ AI Test Selection — PASSED

1 test case(s) selected (view run)

Marketplace image: canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

Count
✅ Passed 0
❌ Failed 0
⏭️ Skipped 1
Total 1
Test case details
Test Case Status Time (s) Message
verify_mshv_crash (lisa_0_0) ⏭️ SKIPPED 0.000 check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.CBLMariner'>] but VM supports [<class 'lisa.o

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Pull request overview

This PR updates the kdump vmcore result-check loop to avoid false failures caused by a fixed ~100s “no growth / no file” stall threshold, by deriving adaptive wait windows from timeout_of_dump_crash and improving failure diagnostics.

Changes:

  • Replaces the fixed max_tries = 20 (≈100s at 5s polling) with adaptive “no file” and “no growth” windows derived from timeout_of_dump_crash.
  • Converts these windows into retry counts (via math.ceil) and surfaces the chosen wait window in raised exceptions.

Comment thread lisa/tools/kdump.py
Comment thread lisa/tools/kdump.py
@lubaihua33 lubaihua33 force-pushed the baihua/kdump-adaptive-timeout branch from 36f0439 to 49829a9 Compare June 30, 2026 09:07
@github-actions

Copy link
Copy Markdown

✅ AI Test Selection — PASSED

1 test case(s) selected (view run)

Marketplace image: canonical 0001-com-ubuntu-server-jammy 22_04-lts-gen2 latest

Count
✅ Passed 0
❌ Failed 0
⏭️ Skipped 1
Total 1
Test case details
Test Case Status Time (s) Message
verify_mshv_crash (lisa_0_0) ⏭️ SKIPPED 0.000 check skipped: OS type mismatch: ["requires [<class 'lisa.operating_system.CBLMariner'>] but VM supports [<class 'lisa.o

@LiliDeng LiliDeng merged commit 290ffd3 into main Jun 30, 2026
61 checks passed
@LiliDeng LiliDeng deleted the baihua/kdump-adaptive-timeout branch June 30, 2026 11:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants