On 12/31/2014 01:57 PM, Kevin O'Gorman wrote:
I've just gotten 4 4TB drives to replace my 4 2TB drives. I'm wanting to
have one normal 4TB drive and one logical 12TB drive, so I will make three
physical drives into one group, one logical volume and one partition
support the big partition. My system actually resides on a fifth: an SSD
drive. I am not interested in RAID, and I'm not sure striping would even
help. I just have gigantic files I need to create and process once in a
while, so it's really temporary space.
mdadm RAID 0 should be the simplest approach for the 12 TB.
I do want to insulate the one drive from any failures on the other three.
That data is not at all temporary, but it is backed up regularly. I want
to limit it's failure profile.
Take a look at ZFS, both for the 12 TB and especially for the 4 TB:
https://en.wikipedia.org/wiki/ZFS
https://packages.debian.org/wheezy/zfs-fuse
http://zfsonlinux.org/
Note that ECC is strongly advised:
https://www.google.com/search?q=zfs+ecc
You can limit hardware failures with ZFS by building pools from
identical pairs of drives.
You can limit software, operator, and certain administrator failures
with ZFS by taking periodic snapshots (e.g. hourly cron job).
You can limit additional types of failures with ZFS by copying and/or
replicating file systems to other devices; notably off-line and/or off-site.
HTH,
David
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/54a5b360.8040...@holgerdanske.com
