Posts

SAP HANA High Availability on AWS with Pacemaker Building a production-ready SAP HANA cluster using Pacemaker, AWS-integrated STONITH fencing, overlay IPs, quorum devices, SAPHanaSR hooks, and preload readiness validation. Figure 1 – Overall architecture showing SAP HANA System Replication, overlay IP resources, STONITH fencing, and a dedicated quorum device. Introduction Implementing SAP HANA High Availability on AWS is not particularly difficult. Implementing it correctly is. ...

How I automated a monthly AMI refresh pipeline using EC2 Image Builder, Lambda, EventBridge Scheduler, launch templates, and warm pools so autoscale instances stopped spending 15–20 minutes patching themselves during scale-out.

A few weeks ago I had to migrate a RHEL 5 server to Hyper-V. Yes, RHEL 5. Released in 2007. Old enough to drive in some states. Naturally, this server was still important. Because of course it was. Nobody keeps legacy servers around because they are fun. They stay around because they run something important, mysterious, undocumented, or all three. So there I was, staring at a server that had survived multiple generations of hardware, operating systems, hypervisors, managers, and probably a few organizational charts. The plan sounded simple: Move the server to Hyper-V. That sentence should come with a warning label. ...

The Problem Nobody Wanted Early in 2026, our team found itself staring at a problem that nobody was particularly excited to own. Bamboo 9.6 — the CI/CD platform used by engineering teams across UC San Diego — was approaching end of life. Atlassian’s recommended path was clear: upgrade to Bamboo 12.1 Data Center. Simple enough. Except for one small detail. The JFrog Artifactory Plugin, used by virtually every team to publish build artifacts, did not work on Bamboo 12.1. ...