Robert Elz <[email protected]> writes: > Date: Mon, 11 May 2015 19:16:42 -0453 > From: "William A. Mahaffey III" <[email protected]> > Message-ID: <[email protected]> > > | I concur on the observation about OS recovery, I want the box to be > | failsafe w/ 1 HDD going out, a *highly* improbable event, > > Not that improbable - they're moving parts, like fans, they wear over time. > Leave the system running long enough, one will die &/or develop bad spots. > > So yes, raid1 (or raid5) everything (including swap space) if you want the > system to keep running for long periods (and of course, depending on what > "keep running" really means, perhaps even arrange for hot swap drives, > so you can replace a bad one without a shutdown.)
Strongly seconded. I have a lot of RAID1 systems between personal and work, and estimating that each disk fails every 3 years is not crazy. I would guess that at least half the systems have had a replacment. With raidframe RAID1, this is not a big deal. Without hot swap, which AFAIK NetBSD doesn't really support well, yes you need to take the machine down. But 10 minutes of planned downtime to put in a new disk -- and no data loss -- within a few days of failure is awesome compared to crash in the middle of the night and scrambling to recover backups.
pgp9fq19MvomK.pgp
Description: PGP signature
