On 03/15/18 12:24, Pascal Hambourg wrote:
Le 15/03/2018 à 06:01, David Christensen a écrit :
2. Instead of RAID1, use a checksumming file system (btrfs), take
images periodically, put key configuration files into a version
control system, and backup data daily. This is what I do for all my
system drives.
All these techniques are backup. RAID does not have the same purpose.
RAID (with redundancy) is designed to provide availability.
Let me add that mine is a SOHO network that does not require 24x7
uptime. And, since switching to solid-state system drives, I have yet
to experience a system drive hardware failure. My most common failure
mode is instability due to system administration mistakes and root
software development bugs. RAID1 won't help me; recovering from an
earlier known-good image will.
Virtual machine snapshots make recovery even easier.
If/when the Debian Installer supports ZFS boot and encrypted root, I can
install ZFS snapshotting software and will have yet another option.
David
