Re: [Nolug] NAS

IISO? Google's not that helpful.

Anyway, I have a large tower case with lots of drives that are
grouped into a single lvm2 LV (logical volume). To back it up
(since it's not yet full) I also have an external USB enclosure with
two 1TB drives that are grouped into an LV. Every so often, I mount
it, execute the appropriate magic to make the PV (physical volumes)
visible and then mount it and do a backup.

In your case, since MPEG2 files (i.e., DVDs) aren't compressible,
I'd mirror the "active" movie folder to the external device with "cp
-avu" or rsync. It'll take a while the first time, but afterwards
it'll only move new/modified files.

"tar cvfj" for your compressible files. Script out something that
will copy (cp or tar, depending on the data) over the files. If
you'd like, I'll send you my script.

This (http://www.directron.com/etcsdu2jbk.html) is the same (or
*very* similar) to what I have.

Once that becomes too much for you, or if you want to Start Big, go
with a 4- or 5-bay eSATA enclosure.

On 2010-04-10 23:05, Shane Russo wrote:
> I think my needs will be more reading. I plan on storing IISO of software
> and pictures and some word docs and such. This is for in home use for
> backups and organization. I have 2 1 TB drives that are quickly filling up
> with DVD backups and home movies. Eventually I would like to build a second
> box and have it replicated offsite somewhere. The problem comes in how do
> you backuo 3-8 TB of data?
>
>
> Thanks,
>
> Shane Russo
>
>
> On Sat, Apr 10, 2010 at 10:54 PM, Ron Johnson <ron.l.johnson@cox.net> wrote:
>
>> On 2010-04-10 22:32, John Souvestre wrote:
>>
>>> Hi Shane.
>>>
>>>
>>> It depends on what you do most – read or write. It also depends on the
>>> size of the data blocks you are reading. For small blocks the disk access
>>> time is important and being able to get all of the data from just one drive
>>> access (Raid 1) is a big advantage. Indeed, with Raid 1 you can get all of
>>> the data from either drive, thus allowing you to overlap reads.
>>>
>>>
>> Note also that there are two kinds of orthogonal reads and writes: length
>> (short, long) and "direction" (sequential and random).
>>
>> BITD (back in the day) RAID5 was faster than RAID1 in long sequential reads
>> because of parallelism (lots of disks and lots of controllers) but RAID1 was
>> faster at short reads and random short. (Random long doesn't really make
>> sense.)
>>
>> Nowadays with capacities so enormous, it's hard to say what the performance
>> of SATA drives would be.
>>

-- 
Dissent is patriotic, remember?
___________________
Nolug mailing list
nolug@nolug.org