Skip to content

Daily: define reboot-safe state for pinned eBPF maps - #213

Open
yunwei37 wants to merge 13 commits into
mainfrom
daily/2026-09-23-ebpf-reboot-state
Open

yunwei37 wants to merge 13 commits into
mainfrom
daily/2026-09-23-ebpf-reboot-state

Conversation

@yunwei37

Copy link
Copy Markdown
Member

Summary

  • publish one new bilingual Daily Report on the boundary between bpffs pin lifetime and reboot-safe eBPF state
  • distinguish live-object structural compatibility from application-state semantics after the old kernel object graph disappears
  • develop per-map reboot state contracts, consistency-aware checkpoint epochs, and staged restore gates
  • refresh the Daily Report indexes plus current SEO/data operating state
  • reconcile the completed September 22 run and preserve the newest-ten mix at 7 eBPF / 1 Agent / 2 adjacent
  • make no unrelated technical SEO change because current evidence does not establish a concrete site defect

Evidence

  • the configured Google Drive folder was rechecked on 2026-09-23; no weekly family newer than 2026-09-14..20 is present
  • finalized observed GSC rows 2026-09-14..19 are 391 clicks / 55,086 impressions / ~0.710% CTR / ~6.79 weighted position versus 376 / 55,036 / ~0.683% / ~6.46 for 2026-09-07..12; complete seven-day and 28-day comparisons remain unavailable because missing rows are not treated as zero
  • newest GA4 weekly aggregate remains partial; Cloudflare remains disabled
  • the 2026-09-23 public-safe brief reports homepage, robots, and sitemap HTTP 200 with 790 sitemap entries, plus current public GitHub/DEV context
  • primary technical evidence comes from current BPF pinning, Linux BTF/map documentation, libbpf APIs, and Cilium production pin-recovery evidence

Editorial boundary

Before this publication the newest ten reports are 7 eBPF-centered / 1 pure Agent / 2 adjacent systems. Today's eBPF-centered report rotates out the eBPF-centered September 9 report, so successful publication preserves 7 / 1 / 2.

This is the fourth boundary in the active eBPF Deployment Compatibility and Lifecycle series. It begins after reboot has destroyed the old kernel object graph, which keeps it distinct from August live transactional upgrade, September 18 cross-kernel program semantics, and September 22 interface negotiation.

Validation and deployment

Final acceptance requires terminal-green final-head CI, complete diff and generated-output self-review, review-thread inspection, squash merge, exact-merge validation and production deployment, bilingual production verification, sitemap verification, and exactly one compact top-level merged-PR closeout comment. No second closeout PR will be created. The recurring operations schedule remains enabled regardless of blockers.

Copilot AI lite review requested due to automatic review settings September 23, 2026 15:48

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Reviewed the head 99e9bea8 as a content change: a new research page
(docs/research/ebpf-pinned-map-reboot-state.md, 283 lines) plus the usual
seo-data bookkeeping (content-series, daily/2026-09-23, plan, status).

Two things I checked because they are the failure modes for this kind of batch:

  • the seo-data edits are additive/diff-only (no whole-file rewrites that would
    drop earlier entries) — status.md and plan.md both change by
    tens of lines rather than being replaced;
  • the research page is English-only. The neighbouring entries (#200, #210, #212)
    each ship a .zh.md twin, so a missing Chinese twin is the one inconsistency
    worth deciding on deliberately rather than by omission.

Checks are green.

AI-generated response; a maintainer will review and follow up later

@yunwei37
yunwei37 force-pushed the daily/2026-09-23-ebpf-reboot-state branch from 99e9bea to d7075b2 Compare November 13, 2026 23:13

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Reviewed at d7075b2.

Nine files: the bilingual report pair docs/research/ebpf-pinned-map-reboot-state{,.zh}.md (283 lines each), index entries in docs/research/index{,.zh}.md, and .github/seo-data/{content-series,daily/2026-09-23,plan,site,status}.md. Front matter follows the same convention as the neighbouring reports (date, slug, quoted title/description, tags led by Daily Report, research_question, source_cutoff, status: daily-report). The route /research/ebpf-pinned-map-reboot-state/ is new, the index insertion sits above the 2026-09-22 entry, and no existing route changes.

The report's factual premise holds: docs.ebpf.io/linux/concepts/pinning/ says map pins are "ephemeral and do not persist over restarts of the system", which is the distinction the report builds on — bpffs preserves a live map across process restart, not across host reboot. The status.md claims about run 35754194946 (Validate SEO Operations) and 35754194965 (Deploy Static App) also hold: both are completed/success at bd751f30. The site.md edits bump the sitemap count 786→790 and refresh star/fork numbers; those are reported values from the publisher, not something I recomputed, and the new routes here account for only part of that delta.

The blocker is the same one as #212: both PRs add .github/seo-data/daily/2026-09-23.md with different content (this one describes the reboot-state report, #212 the map-reuse report) and both rewrite content-series.md and status.md. Both are on base 49cc955 and only look mergeable because neither has landed. Merge one, rebase the other.

Checks at this head: validate and deploy-static-app green, GitGuardian green, sync-external-docs skipped — it skips on every PR in this repo, so it is not evidence of anything.

Verdict: mergeable after the seo-data overlap with #212 is resolved.

AI patrol review

@yunwei37
yunwei37 force-pushed the daily/2026-09-23-ebpf-reboot-state branch 3 times, most recently from df55def to 84664a7 Compare November 15, 2026 08:34

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Reviewed head 84664a78eb: the previous review point (d7075b28ad) moved onto today's main. The branch's own content is unchanged.

Evidence: for every path this pull request changes (9 files), the git blob at this head equals the blob at d7075b28. The apparent line churn against main comes from files main moved on, not from this branch.

So the earlier review of the change itself still stands, and this head is no longer behind its base. Merge readiness is unchanged: it waits on a maintainer's decision.

AI patrol review

@yunwei37

Copy link
Copy Markdown
Member Author

Verdict: ready to merge

Head 84664a78 carries a patrol review of its change (see the review above/below on this pull request), and every check on this head is green. Nothing in the change is left for the patrol to fix.

Merge is the maintainer's to make: eunomia.dev is not in the small-repository group, so the patrol does not merge it.

AI patrol review

@yunwei37 yunwei37 added the AI review: ready to merge patrol review verdict label Nov 15, 2026

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Verdict: ready to merge

The review above stands: this head's content is blob-identical to the previously reviewed revision, so only the rebase onto today's main moved. Checks green, mergeable, level with the base.

AI patrol review

@yunwei37
yunwei37 force-pushed the daily/2026-09-23-ebpf-reboot-state branch from 84664a7 to 42984f9 Compare November 15, 2026 19:08

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Verdict: ready to merge once CI is green

Reviewed head 42984f98. Every changed file is blob-identical to the previously reviewed revision (84664a78): the bilingual ebpf-pinned-map-reboot-state pair, both index edits, and the four .github/seo-data/ files (content-series.md, daily/2026-09-23.md, plan.md, site.md, status.md). Merge base unchanged at f8a2cce8.

Content is unchanged: the report defines what a pinned map's state must be after a reboot for it to be safely re-adopted (map metadata identity, value schema, and the pin lifecycle), and the roadmap-selection rules it adds to the daily record are the same ones reviewed on the earlier head.

Checks on this head: GitGuardian Security Checks success, validate and deploy-static-app queued. The verdict turns to merge-ready when those report green.

AI patrol review

@yunwei37
yunwei37 force-pushed the daily/2026-09-23-ebpf-reboot-state branch from 42984f9 to 8031a25 Compare November 16, 2026 03:27

@yunwei37 yunwei37 left a comment

Copy link
Copy Markdown
Member Author

Choose a reason for hiding this comment

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

Verdict: ready to merge once CI is green

Reviewed at head 8031a251, which is the previously reviewed 42984f98 moved onto today's base. Nothing in the change itself moved: for the range merge-base(main, head)..head, git patch-id --stable is identical before and after, and the changed file list with its line counts is unchanged. The earlier review (5912544833) therefore stands as written, and so does its verdict.

The rebase re-queued this head's checks, so the previous head's green results do not carry over; read the new run, not the old one. Merge stays with a maintainer.

AI patrol review

@yunwei37 yunwei37 added the AI review: merge after CI patrol review verdict label Nov 16, 2026
@yunwei37

Copy link
Copy Markdown
Member Author

Verdict: ready to merge

This is the verdict for head 8031a251. The branch was replayed onto current main; its diff against the base has the same patch-id as the reviewed head, so no line of the change moved — only the base did. All checks on this head are green.

Merge is the maintainer's to make: eunomia.dev is not in the small-repository group, so the patrol does not merge it.

AI patrol review

@yunwei37 yunwei37 removed the AI review: merge after CI patrol review verdict label Nov 16, 2026

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants