LinuxCon + CloudOpen + ELC-E Europe 2014

While robust and generally stable, the Linux operating system may suffer from occasional bugs in the kernel stack, sometimes leading to an irrecoverable condition known as kernel oops. The condition can translate into a system panic, resolved only by a reboot. In time-to-market critical data center environments, kernel crashes can impact the productivity of compute resources. Resolving bugs in the kernel code that cause the panic situations is of paramount importance. In homogenous environments, where a single operating system dominates the install base, individual bugs can potentially manifest on all machines in a very short period of time. In this presentation, Igor Ljubuncic presents on how Intel’s IT transformed its business with an automated Linux kernel crash collection, analysis and reporting infrastructure, designed to address stability of the system’s core – the kernel.