Microcode Mayhem : How a Mainframe Timer Bug Slowed an Entire Development Team

Podcast: IT Horror Stories with Jack Smith · Jack Smith

Opublikowano: 2026-08-17

One day, a busy IBM-compatible mainframe began slowing to a crawl. Hundreds of programmers noticed immediately—some even resorted to banging on the server room window whenever the system became unusable. The problem appeared after a microcode upgrade, but nothing pointed to an obvious cause. In this episode, Jack Smith talks to Matt about an extraordinary piece of low-level troubleshooting that led deep into the internals of VM, interrupt handling and timer queues. Armed with nothing more than hex dumps, careful reasoning and a willingness to modify live memory, Matt uncovered a vendor bug and devised a te...