Patchy mirror upgrade
patchy mirror upgrade¶
Move pins to newer upstream versions and regenerate derived state
Synopsis¶
Converge entries onto their target versions: resolve the newest upstream version satisfying each entry's constraint, splice the pin, re-pick tracked image tags past their cooldown, re-vendor, re-render, and regenerate the digest locks. A fully-converged entry is a clean no-op.
Entries sharing a lockstep group bump together: the group target is the lowest of the members' newest satisfying versions, holding the bump until upstream has published every member's tag.
--check reports the plan as data without touching the tree (exit 0 with an empty groups list when everything is current). patchy never commits: the calling pipeline turns the mutated tree into branches and PRs.
Examples¶
patchy mirror upgrade --check -o json
patchy mirror upgrade --all
patchy mirror upgrade --group flux --to 0.57.0
patchy mirror upgrade opentelemetry-collector
Options¶
--all upgrade every entry in the store
--check report the update plan without touching the tree
--group string upgrade one lockstep group
-h, --help help for upgrade
--to string target version (default: newest satisfying the constraint)
Options inherited from parent commands¶
-A, --all-namespaces work across every namespace
--context string kubeconfig context to use
-C, --directory string mirror store directory (walks up to mirror.yaml; default: the working directory)
--kubeconfig string path to the kubeconfig file
-n, --namespace string namespace to work in (default: the context's)
--no-color disable colour and styling
-o, --output string output format: table, wide, json, yaml, name, or markdown (default "table")
--request-timeout duration timeout for a single API call (default 30s)
-v, --verbose log what the CLI is doing to stderr
--workers int concurrent registry operations per stage (default 4)
SEE ALSO¶
- patchy mirror - Mirror upstream helm charts and OCI artifacts into a platform registry