naturerecordists
[Top] [All Lists]

2. Re: So, I have lost everything...

Subject: 2. Re: So, I have lost everything...
From: "Eric Fassbender" eric.fassbender
Date: Wed Apr 3, 2013 3:53 am ((PDT))
Hi  Kin,

I agree that there is still the element of the RAID server that can fail an=
d in your case it has happened, so apparently it is not unlikely. So I real=
ly like your solution/approach. Am I assuming correctly that you use one dr=
ive that always stays the primary and that then gets backed up to the two o=
ther external drives? If so, what software are you using to perform this fe=
at or do you do it manually? I would be very interested to hear your feedba=
ck as although I am still confident in my RAID, it has the limitation that =
it is slowed down by its only I/O connection - a 1GBit Ethernet connection.=
 So I was eyeing a USB3 RAID solution using 2 or 3 standard external HDD an=
d SoftRAID 4.5 (Mac), however, I am wondering whether your solution offers =
any benefits over a software RAID?

For anyone who is not that familiar with RAID systems, there are different =
types of RAID. What we need for data protection/safety is RAID 1, a 'Mirror=
ed Set'. This can be 2 or 3 even 4 drives and they all hold the same mirror=
ed data. So always check the specifications because there are many RAID sol=
utions out there that offer RAID 0, which makes data transfer super fast bu=
t does not offer protection against data loss. So if you are in the market,=
 look for RAID 1.

Cheers,

Eric


On 03/04/2013, at 7:49 PM, Kin Corning <> wrote:

> I would offer some comments about using desktop RAID systems, which have =
been mentioned in this thread.
>
> When RAID arrays are used in Storage Area Networks or other such configur=
ations in professional datacenters, *everything* is redundant -- the server=
s, the disk storage, power supplies, RAID controllers, and the interconnect=
ions between server and disk. In a desktop RAID unit, typically it is only =
the disk drives themselves which are protected. And, while a disk is the pa=
rt most likely to fail, other failures can occur as well.
>
> I was reminded of these limitations the hard way a couple years ago when =
I had a desktop RAID unit fail, apparently (I'll never know for sure) due t=
o a failed controller board in the box. The device was (just) out of warran=
ty, the manufacturer (a major well-known firm) would not offer any servicin=
g, and their only suggestion was to buy another identical unit, transfer th=
e drives, and see if the data was intact. I did a fair bit of exploration a=
nd could not find any third-party who had the expertise to service the unit=
, which became an expensive scrap job.
>
> I was backed up separately, so my data was protected, but out of curiosit=
y I contacted one of the best known firms in the UK for a quote on data res=
toration, and the answer was =A33000 to give it a try with no guarantee of =
success. This may sound crazy but is understandable given the complexity of=
 how data is distributed across drives in the various RAID strategies, and =
the fact that the data may or may not have been corrupted.
>
> These desktop RAID units may still be attractive to some in offering a de=
gree of real-time protection against a drive failure, but in my own setup (=
used primarily for photography) I have concluded this benefit isn't worth t=
he expense and complexity. I have moved to a simpler system with 4 identica=
l large-capacity drives, 3 of which are onsite and 1 of which is always off=
site. Every day (twice a day if I am doing a lot of work) the primary drive=
 gets backed up to each of the 2 other onsite drives, and as often as is pr=
actical one is swapped with the offsite device, to keep the data protected =
from fire or burglary as up-to-date as possible. In this setup, I may lose =
a few hours of work if a drive fails, but on the other hand the system is i=
nstantly restorable and the cost of replacing the failed unit is much lower=
.
>
> Not to state what may be obvious, but if you choose to use one of these u=
nits notwithstanding, it is important to remember it is not a backup device=
. RAID offers real-time protection against a drive failure, but an accident=
ally-deleted file will be deleted across the RAID array, and any software-g=
enerated file corruption likewise may be duplicated in each 'copy' of the f=
ile. So even if you employ one of these units to protect your work intra-da=
y, you still need to back up to separate devices regularly.
>
> Kin
>
> --- In  "sounds.images" <=
..> wrote:
> >
> >
> > Many thanks everyone for you words of support and replies, a new laptop=
 has been ordered and I do have a caddy already, so will put it in ther and=
 have a go at transferring it of it is possible to recover them..
> >
> > Seen in PC world.co.uk is a USB 3.0 1tb external drive for =A359.00 so =
may grab one..
> >
> > Will update as soon as I can.
> >
> > Simon
> >
> >
> >
> >
> > --- In  "sounds.images" <sounds.images=
@> wrote:
> > >
> > > Last night the other half shut down the laptop as normal before bed. =
This morning it was dead. It would power on and fans would spin but nothing=
. Took it to my parents and tried their hard drive in my machine.... Nothin=
g..
> > >
> > > Tried my hd on their machine and it tried to boot but went straight t=
o blue screen of death twice...
> > >
> > > So never put off your back ups. I have lost every recording I made si=
nce September. It was on the cards to do this little job this weekend, but =
it is too late.. Will try again when I get a replacement machine. This was =
completely unforeseen..
> > >
> > > Oh well... Lesson learned. Never put off the most important job of al=
l.
> > >
> > > Simon.
> > >
> >
>
>










<Prev in Thread] Current Thread [Next in Thread>
  • 2. Re: So, I have lost everything..., Eric Fassbender <=
Admin

The University of NSW School of Computer and Engineering takes no responsibility for the contents of this archive. It is purely a compilation of material sent by many people to the naturerecordists mailing list. It has not been checked for accuracy nor its content verified in any way. If you wish to get material removed from the archive or have other queries about the archive e-mail Andrew Taylor at this address: andrewt@cse.unsw.EDU.AU