budctl 0.3.1: stop reporting false blockers on clusters that work - #1
Merged
Merged
Conversation
A readiness run against a k3s cluster Bud was already serving reported three blockers, and none of them was stopping anything. This corrects the checks and the catalogue behind each false finding, and keeps the real ones. storage - csi-healthy matched every rancher/* image to rancher.io/local-path, so a Pending k3s ServiceLB pod read as a broken storage driver. The vendor label is now a last resort, used only when the provisioner has no more specific name (driver.longhorn.io, openebs.io/local). - csi-healthy no longer reports "registered on 0/N nodes" for a provisioner that is not a CSI driver. - expansion no longer asks node-local classes for allowVolumeExpansion: local-path ignores the requested size and has no resizer, so the node's disk is the ceiling and the suggested patch changed nothing. egress and charts - The prometheus-community, bitnami, CloudNativePG, Altinity, Percona, SeaweedFS and OpenTelemetry repos are optional: every ApplicationSet installs those charts from registry.bud.studio as packaged OCI charts with the subcharts inside, so the cluster never fetches them. dapr stays install-time. - Let's Encrypt moves to a new egress.acme check, skipped unless TLS is obtained through ACME, and it states the real consequence (no certificate) instead of "the sync stops at the first image or chart". - egress.proxy no longer treats k3s's NO_PROXY on helm-install pods as a proxy. argocd - chart-repo-credential tries an anonymous pull before predicting a 401: a charts project served to anonymous clients needs no repository Secret. gpu - The functional probe defaults to debian:trixie-slim. HAMi preloads its vGPU library into every GPU container and it needs libdl.so.2, so busybox's shell died in the loader. A loader error is now named, with the flag to change it. registry - budimages.azurecr.io is retired and removed from the inventory. report - Reports saved from the interactive view carry the budctl version, catalogue, cluster and domain. Report.Stamp is shared by both output paths; the TUI used to save before main stamped the report, so the file said "cluster unreachable". Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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
1 participant
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.
Why
A
budctl checkagainst the tcs-vmware k3s cluster, which Bud was already serving, reported three blockers. None of them was stopping anything. This PR fixes the checks and catalogue entries behind each false finding, keeps the real ones, and bumps the installer pin to 0.3.1 so the release can be tagged frommain.storage.csi-healthyrancher.io/local-pathmatched everyrancher/*k3s image, so a Pending ServiceLB pod counted as a broken driveregress.install: Altinity chartsegress.install: Let's Encryptegress.acme, skipped unless TLS uses ACMEregistry.from-cluster: azurecrargocd.chart-repo-credentialstorage.expansionallowVolumeExpansion, but local-path ignores sizes and has no resizergpu.operator-functionallibdl.so.2; busybox's shell died in the loaderdebian:trixie-slim; loader errors are namedmainstamped the reportReport.StampUnchanged, because they are real:
registry.from-clusterforregistry.cn-hangzhou.aliyuncs.com(HAMi's default scheduler image), the Helm/ArgoCD ownership overlap on argocd, and 2 nodes where the data stores expect 3.Other changes
egress.proxyno longer treats k3s'sNO_PROXYon helm-install pods as evidence of a proxy.charts.classicwording updated now that only dapr is sourced from an upstream repo.2026-09-15;install.shpins0.3.1.Verification
go build,go vet,go test ./...(including the failure-coverage gate), cross-compiles for all four platforms, andsh -n install.sh.install.shpins 0.3.1, and a binary built with that version reportsbudctl 0.3.1.HEAD bud:1.2.8returned ok, and the GPU probe printedBUDCTL_DEVICE_PRESENT.Release
After merge, tag the merge commit
v0.3.1and push the tag.release.ymlthen tests, builds and publishes.🤖 Generated with Claude Code