nickboldt

# it never ends

oc get snapshot -n rhdh-tenant -o json | \
  jq -r '.items[].metadata.name | select(startswith("rhdh-plugin-catalog-2-2026"))' | \
  xargs -r oc delete snapshot -n rhdh-tenant
oc get snapshot -n rhdh-tenant -o json | \
  jq -r '.items[].metadata.name | select(startswith("rhdh-plugin-catalog-1-202"))' | \
  xargs -r oc delete snapshot -n rhdh-tenant

# etc.

nickboldt

#weekly maintenance

glab ci run -R rhidp/rhdh -b main --variables "WEEKLY_MAINTENANCE:true"
brave-browser https://gitlab.cee.redhat.com/rhidp/rhdh/-/pipelines?source=api&ref=main&scope=all

# update plugin workspaces (maybe, if changes found)

branch=main 
branch=release-2.1

nickboldt

uv run scripts/trigger_nightly_job.py \
  --job periodic-ci-redhat-developer-rhdh-plugin-export-overlays-release-1.10-e2e-ocp-helm-nightly \
  --catalog-index-image registry.access.redhat.com/rhdh/plugin-catalog-index:1.10.4-1788447603

nickboldt

The GitLab job did not fail on missing types or a broken yarn.lock chain. Loop 1 exported successfully. Loop 3 then failed yarn.lock equivalence for two non-embedded frontend libraries.

What failed

Job sync-midstream #60069298:

• orchestrator:validation-failed
• [DRIFT] BODY DRIFT for @red-hat-developer-hub/backstage-plugin-orchestrator-form-api@2.10.0 and ...-form-react@2.11.0 on orchestrator and orchestrator-form-widgets
• Backend plugins that already --embed-package orchestrator-common / orchestrator-node passed
• ssh2 / cpu-features gyp errors are optional native rebuilds (Failed to build optional crypto binding) and are not the job failure

nickboldt

Skip e2e for base-image bot PRs (path skip + working /override)

Draft plan for discussion. Overview: stop Hive/cluster claim for base-image-only PRs via skip_if_only_changed in openshift/release, and make /override ci/prow/e2e-ocp-helm succeed by posting as rhdh-bot (via RHDH_BOT_TOKEN) after authorizing that user in top-level OWNERS.

Context

Today two things fight each other:

1. [skip-e2e] is too late — checked inside redhat-developer-rhdh-ocp-helm-commands.sh after cluster_claim.

nickboldt

oc get pipelineruns \
-n rhdh-tenant -l 'appstudio.openshift.io/application=rhdh-plugin-catalog-1' -o json | \
jq -r '.items[] | select(.status.conditions[0].reason == "Running" or .status.conditions[0].reason == "PipelineRunPending") | select(.metadata.name | test("-on-pull-")) | .metadata.name' \
| xargs -r -n 25 -P 4 sh -c 'for pr; do tkn pipelinerun cancel "$pr" -n rhdh-tenant 2>/dev/null || \
oc delete pipelinerun "$pr" -n rhdh-tenant --wait=false 2>/dev/null; done'

nickboldt

alias lgtm='gh pr review  -a -b "/lgtm
/approve" '

alias lgtm-override='ght;gh pr comment -b "/override \"ci/prow/e2e-ocp-helm\"
/override \"Build with Node.js\" 
/override \"Test with Node.js\"
/override \"Build Image\" "'

alias merge='gh pr merge -s --auto '

nickboldt

# Close JIRAs recursively

use `jira` commandline to close Feature <link> with comment "Closing as won't do."; 
then look at child Epic tasks under that issue and close them too, with the same comment; 
recurse into any children of those Epics and close them using the same comment too.

nickboldt

alias cur='sudo ulimit -v 62914560 && cursor'

other approaches: https://search.brave.com/search?q=limit+the+ram+use+by+cursor+on+lnux&summary=1&conversation=08f722cccea181c2fa2ce5e51198b466f648

nickboldt

# log in to registry
RASRC_REGISTRY="registry.***"
RASRC_USER="***"
RASRC_PWD="**********"
echo -n "[INFO]: Log into $RASRC_REGISTRY ... "
echo "${RASRC_PWD}" | skopeo login -u="${RASRC_USER}" --password-stdin ${RASRC_REGISTRY}

# run konflux push
./build/scripts/kfuxRelease.sh -v 1.9.0 --plugin --stage --debug --regex orchestrator