Kernel panic and oops

In short: A kernel panic is the Linux kernel crashing on an error it cannot recover from. An “oops” is a less severe error after which the system sometimes keeps running.

In more detail: The counterpart of the blue screen on Windows, see BSOD Kernel Security Check Failure. The kernel writes a message with call stack and registers.

In detail

• Causes: faulty memory, faulty drivers (often proprietary modules), broken file system, wrong kernel, hardware faults.
• Typical messages: Kernel panic – not syncing: VFS: Unable to mount root fs, BUG: unable to handle page fault.
• Searching: journalctl -k -b -1 shows kernel messages of the previous boot, dmesg.
• Surviving: with kdump a memory image is written for analysis.
• Tests: memtest86+ for RAM, smartctl for drives.
• Fallback: choose an older kernel in the boot menu.
• Configuration: kernel.panic=10 reboots 10 seconds after a panic.

See also: Kernel modules and drivers, Boot process: UEFI, GRUB, initramfs and kernel