[NCLUG] Installing ubuntu 12.04 to ZFS root directory

[Zak Smith]

I believe the following commands would give a relevent status report
on the pools:

zpool status -x
zpool status

(from this guide
http://docs.oracle.com/cd/E19082-01/817-2271/gbbuw/index.html )

Combined with running weekly zfs scrubs and smartd, the latter of
which will report SMART status per /etc/smartd.conf, you should stay
apprised of impending problems.

Just a note on the redundancy- I would rather have two physically
separate machines (one primary, one backup) with less drive redundancy
per machine and regular (daily or hourly) checkpoints between them
than extra extra redundancy inside my primary working machine and no
secondary machine backup. 

I did work up a set of scripts that send the incremental checkpoint
data from ZFS to a remote ZFS filesystem.  I plan to package it up
with an explanation and post it one of these days.



On Thu, Aug 22, 2013 at 08:15:52AM -0600, Jesse Griffin wrote:
> Of course, I agreed with your crazy assessment out of jest.
>
> You are correct that any 2 drives in your 4 drive RAIDZ2 could fail and you
> would be okay.  Your mathematical calculations are interesting.
>
> I always endeavor to have a spare drive for any RAID setup that I have, so that
> I can address a failed drive as soon as possible.  A CRON or some sort of
> monitoring system can ensure that you are made aware of zpool problems right
> away too.
>
> You do have backups of your /home partition on a completely different filesystem
> and/or computer, right?



--
# Zak Smith    mobile 970-232-4468