Commit c071075
tazhate
fix: webhook path, CRD chain enum sync, batch script exit code
Three real bugs surfaced by the manual batch e2e workflow:
1. Webhook path mismatch — kubebuilder marker on ChainInstanceValidator
still pointed at /validate-nodes-chainplane-io-... from the old
nodes.k8s-bch.io group, so admission requests for chains routed by
ValidatingWebhookConfiguration to /validate-chains-... got "the
server could not find the requested resource". Renamed marker to
match the new chains.chainplane.io group and regenerated
config/webhook/manifests.yaml plus the chart's webhook template.
2. Chart CRD was out of sync — charts/chainplane/templates/crds/
chains.chainplane.io_chaininstances.yaml shipped an enum of 18
chains while the Go type lists 95. Re-vendored both
chaininstances.yaml and chainversioncatalogs.yaml from
config/crd/bases/ so the chart's enum matches the source of truth.
3. scripts/manual-batch-test.sh tallied failures inside a `{ ... } |
tee` subshell, so `fail` was always 0 in the parent and the script
exited 0 even when every batch failed. Split tee out and count
failures in the parent shell.
Context: gh run 25135219989 (the first sweep of all 103 samples) came
back conclusion=success despite 103 FAIL rows in summary.txt. Pulled
the artifact, grepped logs, found "Unsupported value" (CRD enum) and
"failed calling webhook" (path), then traced the script bug from the
exit-code anomaly. ~30min of investigation, ~10min of fixes.1 parent 19ba7ad commit c071075
5 files changed
Lines changed: 204 additions & 69 deletions
File tree
- api/v1alpha2
- charts/chainplane/templates
- crds
- config/webhook
- scripts
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
197 | 197 | | |
198 | 198 | | |
199 | 199 | | |
200 | | - | |
| 200 | + | |
201 | 201 | | |
202 | 202 | | |
203 | 203 | | |
| |||
0 commit comments