Mitigating the impact of hardware failure
Software can have bugs, and PostgreSQL is no exception. Bugs in the database software rarely, if ever, lead directly to data corruption. Hardware can fail too, but hardware problems are not always so straightforward.
Disk, CPU, or memory failures don't always cause the server to crash. In fact, these failures can persist for weeks or even months before their detection by a monitoring infrastructure. Disk failures are generally abstracted away by RAID or SAN devices, and these arrays are designed to readily handle online rebuilds. Other types of failures are more subtle.
CPU or memory problems can manifest in several different ways. In order for PostgreSQL to function, the data from disk must be read into ...
Get PostgreSQL 9 High Availability Cookbook now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
