[lhotari]

Motivation

Some GitHub runners have less disk space and fewer CPUs than others. As with #26558, one goal is to make Pulsar CI work on smaller runners.

• On runners with fewer CPUs, the unit test group BROKER_GROUP_1 can go over the default 60-minute job timeout.
• The BACKWARDS_COMPAT integration test group pulls several older apachepulsar/pulsar images in addition to the java-test-image, so it can run out of disk space on runners with less free disk.
• Each job keeps the downloaded gradle-build-outputs.tar and Docker image archives on disk after extracting or loading them. This wastes disk space for the rest of the job.

Modifications

• Set the BROKER_GROUP_1 unit test timeout to 70 minutes, 10 minutes more than the default of 60.
• Run the clean-disk action for the BACKWARDS_COMPAT integration test group. It uses the default mode with clean-threshold-gb: '30', not full mode. The action does nothing if 30 GB or more is free. Otherwise it removes /usr/local/lib/android and then /opt/ghc until the threshold is met, and falls back to cleaning apt state.
• Remove gradle-build-outputs.tar in the same step that extracts it, in every job. Remove the downloaded Docker image archive in the same step that loads it, in the integration test and system test jobs. Each step stops at the first failing command, so the archive is only removed after a successful extract or load.

Verifying this change

• Make sure that the change passes the CI checks.

This change is a trivial rework / code cleanup without any test coverage.

Does this pull request potentially affect one of the following parts:

If the box was checked, please highlight the changes

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment