docs(sqlserver,troubleshoot): sediment v18 PR #10265 SSA conflict verification + sideload CRD-set match guide#247
Open
weicao wants to merge 2 commits into
Open
Conversation
…ification + sideload CRD-set match guide - sqlserver-helm-reinstall-ssa-field-manager-conflict-case.md: replace the open verification plan section with the actual PR #10265 verification record (PASS), including 4-field patch image identity, before/after managedFields verdict tables (manager=kubeblocks no longer owns f:spec.* on cmpv + 3 mssql-dynamic-* OpsDef), helm upgrade rc=0 in 5s, and the PR #10264 main-HEAD failed attempt with CRD set mismatch as first-blocker. - New troubleshoot/addon-sideload-crd-set-must-match-vcluster-release-guide.md: sideload controller patch image must match vcluster release CRD set; main HEAD image into release-N vcluster causes the controller starts halfway symptom (Running but missing workers / no matches for kind). Covers selection rule, boot evidence checks, behavioral fallback, anti-pattern reference, and a 5-question self-check. - troubleshoot/README.md: index the new guide. Boundary kept: PR head sideload validation is mechanism-level; the merge-later official image still needs re-verification with the same 4-object managedFields rubric.
Contributor
Author
|
HOLD for docs gate. The new troubleshoot guide topic is useful and the public hygiene is clean, but this branch needs one cleanup round before merge:
Already checked: commit messages and PR body are public-hygiene clean; the new guide is 114 lines and has the right engine-neutral shape. |
…247 review Apply 4 cleanups per controller team review feedback: - "Lily / Rocco" agent attributions → "KubeBlocks 控制器团队" - "Rocco 二审 LGTM (msg=5256518c)" → "二审要点" (drop Slock msg ID) - "Lily 的 PR" → "KB controller fix" - Slock agent workspace paths (with agent UUID) → descriptive "sqlserver mix-A-1 lane sideload-pr10265 evidence bundle" reference Doc readers external to the SQL Server lane cannot access agent workspace paths or resolve internal Slock msg IDs. Mechanism description and 4-obj managedFields verdict tables are unchanged. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Contributor
Author
|
HOLD remains for docs/public-hygiene gate. The useful part is still clear: the new CRD-set sideload guide is the right topic and the PR #10265 mechanism evidence belongs in docs. But the branch is not mergeable yet:
Once these are fixed, I can re-run the merge gate. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
docs/cases/sqlserver/sqlserver-helm-reinstall-ssa-field-manager-conflict-case.md— replace the open verification plan section with the actual PR #10265 verification record (PASS): 4-field patch image identity, helm install+upgrade rc=0 5s, before/after managedFields verdict tables showingmanager=kubeblocksno longer ownsf:spec.*on cmpv/mssql + 3 mssql-dynamic-* OpsDef. Also records the PR #10264 main-HEAD first-blocker (CRD set mismatch on release-1.0 vcluster).docs/troubleshoot/addon-sideload-crd-set-must-match-vcluster-release-guide.md(new) — engine-neutral guide for the sideload-base-branch trap: main HEAD controller image into release-N vcluster makes controller appear Running but lose half its workers (no matches for kindfor CRDs not present in release-N). Covers selection rule, boot evidence triage, behavioral validation, comparison with image-selected-correctly path, anti-pattern reference, and self-check.docs/troubleshoot/README.md— index the new guide.Why
Mix-A-1 lane (SQL Server addon) v18 SSA field-manager conflict surfaced a KubeBlocks controller defect. Lily and Rocco shipped PR #10264 (main) and PR #10265 (release-1.0 backport). On 2026-05-19 the SQL Server team built/sideloaded the PR #10265 head patch image into an idc2 release-1.0 vcluster and verified the mechanism: all 4 target objects now have
manager=kubeblocksowning ZEROf:spec.*subfields, helm in-place upgrade 1.0.4-with-fix1268 → 1.0.5 completes rc=0 in 5s with no SSA conflict.The first attempt with PR #10264 main HEAD failed not because of the fix mechanism but because main HEAD controller binary requires CRDs (Rollout / Instance / ParameterView) not present in release-1.0 vcluster — a CRD set must match vcluster release trap that deserves its own troubleshoot doc so other addon teams sideloading controller patch images don't repeat the same investigation.
Boundary
HandleCRDeletionhas 10+ other callers using the same mechanism; this case did not enumerate them.Test plan