Hi all, I’d like to start a discussion on setting up scheduled CI workflows to regularly build Apache Cloudberry (both the 2.0.0 branch and the main branch) and deploy demo clusters on multiple operating systems — for example, Ubuntu 24.04 LTS, Rocky Linux 8, and Rocky Linux 10.
Currently, our PRs are tested using a development image based on Rocky Linux 9, and @Leonid is also working on enabling Ubuntu 22.04 support in the workflow (see PR #1378, which needs further review and approval — your feedback is welcome!). However, other popular enterprise Linux OS releases are not yet included in our automated workflows. While these systems share most of the same build and dependency steps, I’ve encountered multiple package differences during manual tests on Rocky Linux 10 and Ubuntu 24.04 that caused build failures. In addition to dependency inconsistencies, there are also issues caused by higher gcc/g++ version standards or differences in Python environment handling, which can affect compilation behavior. If we could schedule weekly builds (e.g., every Monday UTC 00:00) for both the Cloudberry 2.0.0 and main branches across these OS environments — including full compilation and demo cluster setup (more tests are better)— it would greatly help us detect issues earlier, particularly those related to compiler or dependency version mismatches. Looking forward to your thoughts! Best, Dianjin Wang --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
