When a Unix process crashes badly enough — segfault, illegal instruction, unhandled signal — the kernel writes a snapshot of the process’s entire memory to disk: a core dump. The name comes from magnetic core memory, the dominant RAM technology of the 1950s and 1960s, making “core dump” yet another piece of Unix vocabulary haunted by hardware that no living sysadmin has ever touched. Core files are invaluable for post-mortem debugging and also reliably fill up /var on production servers at 3am. Many systems now ship with core dumps disabled by default to prevent sensitive memory contents landing on disk, which is responsible, and means that when something does crash spectacularly, you have exactly nothing to debug with.
