You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Document what a new version can refuse, and bump to 1.13.0 (#25)
Three releases in a row have tightened what the config accepts, and until
now the only place that was written down was the error message itself. A new
docs/upgrading.md answers each one: the dry run to do first, what actually
happens per deployment if you skip it, and then a section per refusal
headed by the verbatim message so that pasting a log line finds it.
The seven messages in that page were generated by running -check against a
config that trips each one, not copied from the source. A page like this is
worthless if a character drifts.
The deployment section is the part worth reading twice. A running instance
that reloads keeps its pages. A Kubernetes rollout stalls on 503 readiness
and the old pod keeps serving. Compose restarted comes back healthy on
liveness and serves the getting-started page instead of the site, which is
the one case the dry run exists for, so docker-compose.md now says so where
the image tags are chosen.
Linked from the README table, the docs index and llms.txt, and every pinned
1.12.0 in the chart and the deployment pages moves to 1.13.0.
| **Start** | [Getting started](docs/getting-started.md), the five-minute path |
170
+
| **Start** | [Getting started](docs/getting-started.md), the five-minute path · [Upgrading](docs/upgrading.md), what a new version can refuse and how to fix it |
0 commit comments