zookage

Imported from zookage/zookage (AGENTS.md). Install upstream with npx skills add zookage/zookage. Copyright stays with the author.

AGENTS.md

Purpose

This document defines how AI coding agents should be used safely and consistently in this repository, covering both contributor workflows and end-user agent usage.

Working Principles

• Keep changes minimal and scoped to the request.
• Prefer repository scripts and established conventions over custom workflows.
• Do not run destructive git/history actions unless explicitly requested.
• Preserve existing style and license-header expectations.

Guidance for Repository Users

• Run ./bin/up to spin up a Hadoop cluster; it builds manifests with the ConfigMap template Transformer in a local zookage-util Docker container before applying them, without saving evaluated configuration files locally. Set ZOOKAGE_UTIL_IMAGE to select a locally built renderer image.
• Run ./bin/down to shut down the cluster
• Run ./bin/kubectl <args...> to run kubectl in the zookage namespace
• Run ./bin/logs [<pod name>] to tail logs from all pods or a specific pod
• Run ./bin/ssh [<pod name>] [<container name>] to log in to a container (defaults to client-node-0)

Guidance for Developers

• Primary quality gate: ./test/lint.sh
• Focused lint checks (same tools as ./test/lint.sh): ./test/hadolint.sh, ./test/shellcheck.sh, ./test/yamllint.sh, ./test/license.sh
• To run the console module formatting checks and tests: mvn -f console/pom.xml test
• To auto-fix license headers: ./test/licensefix.sh
• To configure local Docker image builds, copy docker/sample-build.env to docker/build.env and adjust source directories/image settings.
• To build a Docker image, run the matching docker/build-<component>.sh script (for example, docker/build-console.sh or docker/build-hadoop.sh).
• Before dependency bumps, run ./.agents/skills/update-dependencies/scripts/find_dependency_refs.sh [pattern] to find repo-local pins and mirrored consumers.
• Integration test suite on the current Kubernetes profile: ./test/integration.sh
• To switch Kubernetes profile and run the integration suite: ./test/integration.sh [all|auth|ha|llap] (copies test/kubernetes/<name>.yaml to kubernetes/kustomization.yaml and restarts the cluster)
• For full pre-merge validation across every checked-in Kubernetes profile, use ./.agents/skills/final-validation/scripts/run_all_variants.sh (--list, --dry-run, --continue-on-failure, and --no-restore-cluster are available).
• To run an individual integration check directly: ./test/integration/<name>.sh (for example, ./test/integration/web.sh, ./test/integration/spark_sql.sh, or ./test/integration/container_warn.sh)
• To execute a one-off command in client-node-0 during integration/debugging: ./test/integration/run.sh <args...>
• S3-backed coverage is embedded in the MR, Spark, Spark SQL, Hive on Tez, Ozone, and Trino integration checks through test/integration/s3.sh; run the relevant subsystem check rather than treating it as a standalone suite.
• To inspect or regenerate checked-in JCEKS secrets, use ./test/jceks show <jceks-file> or ./test/jceks create <jceks-file>; create reads key-value pairs from <jceks-file>.properties.
• After changing Ranger Kerberos principals or passwords in kubernetes/base/common/config/openldap/bootstrap.ldif, run ./docker/generate-keytabs.sh to regenerate kubernetes/base/common/secret/ranger/ranger.keytab.