Back-up solution - new motherboard?

[just_laze]

Hi guys,

I've just invested in two 1TB drives, one to store my media and the other to back it up to. I tried to perform my first backup last night, 229GB of data, using Cobian. Come the morning the machine wasn't very happy at all, the backup hadn't completed (was stuck on 7&#37 and after I cancelled the backup the machine froze solid. It was freezing on and off beforehand (as I attempted to cancel the back-up).

I've since ran Orthos for 8 hours and it reported no errors. I'm beginning to think the chipset wasn't happy with Cobian running from the system disk (reading/writing), the backup being taken from disk 1 (reading) and placed onto disk 2 (writing) while they're all connected to the same SATA host controller. Do you think this is a plausible explanation for the failed backup?

What I'm now considering is purchasing a motherboard with an onboard RAID 1 controller and would like some suggestions please (I wouldn't mind a little bit of overclocking too!). I've got £100 to spend, not a penny more and my current spec includes:

• E6300
• 2GB DDR2 667
• 9600GT

Do you think employing RAID 1 as the solution is the sensible direction to take? What are my alternatives to buying a new motherboard, if any? I did consider connecting one of the drives via USB (to free up some bandwidth on the SATA controller) instead of e-SATA but that seems so... lacking in elegance and more of a bodge-job than a solution, heh.

Your knowledge and advice is welcome!

laze.

[peterb]

RAID 1 isn't really a back-up solution. True you have two drives which are mirrored, but if the machine fails (say a PSU surge) or the hard drive controller fails, or (as happened to me) a memory failure causes duff data to be written - it will be written to both drives. RAID 1 is all about resiliance and uptime.

For backup the storage medium should be off line for most of the time - except when a backup or restore is being made.

That isn't to say that using your second !T drive cant be used.

So options might be

1. get a removeable draw for the second drive. Set up RAID 1 but once the raid is built, remove the second drive. Clumsy and clunky IMHO!

2. Get the second drive - in a removeable draw as before - and then take an exact clone of the working drive from time to time (Weekly - monthly - depends on how frequently your data changes and how much ypu can afford to lose. The advantage there is that if the working hard drive fails, you have a backup ready to slot in and go - no rebuilding the OS etc. But cloning a 1 T hard drive will take quite a while.

3. Install the second drive in a USB or Esata caddy (Esata will be faster) and connect it to the PC as an when needed. Keep your data on a separate partition and just copy that partition to the backup drive as required.

One thought - of course while you are writing to the hard drive -should you get a failure, you have lost the backup - and the live data - so you should really have two drives (or two 500Mb Partitions) and back up to each one alternately. two partitions limits the amount of data to 500Mb - otherwise you are looking at 2 1T drives and two enclosures.

At the4 end of the day it comes down to risk assessment. When will it fail - and what can you not afford to lose at any price.

Remember that there are two types of computer owner - one whose hard drive has failed - and the one where it hasn't - yet. Drives being mechanical will fail - the only unknown is when. (and that goes for the backup drives too)

Unless you really need a new mobo - IMHO your your £100 would be better spent on the drives.

[2Cold Scorpio]

I pretty well agree with peterb. I'd say an an enclosure to make that second drive an external hard drive (that you can disconnect when backing up is done. Do it, say once every few days to once a week). This way the backup is isolated from the computer if something happens to it (it would only be in danger while the backup occurs, but this is unavoidable). Also, if the backup drive fails separately, you still have your data on the primary drive (one would hope), so once you got a new backup drive, you could resume your archiving.

[just_laze]

Right,

Thanks for your replies guys.

Discarding the option of RAID 1 for the reasons stated above both drives are in external enclosures and are connected via e-SATA. I'm curious as to why the back-up failed/computer died. Do you think it was because the controller was swamped with too much data, maybe the south-bridge over-heated (I was in bed, I wouldn't have known)? Why did the machine become unstable... I guess that's the real question. Bah!

Will run some more tests to see if I can get to the bottom of this.

laze.

[peterb]

I wouldn't have thought the controller was swamped - the throughput would just drop. Overheating is a possibility.

[just_laze]

When you say drop... Oh, the speed of the transfer decrease, yes that would make sense.

Perhaps the problem lies with Cobian? The software back-up solution I used? Maybe I should just try doing it manually, good old fashioned copy and paste. That wouldn't be ideal though because copying and pasting 229GB is something I'd want to leave going overnight and the lack of "shutdown when finished" would be a shame/annoyance.

Is there anything else out there even nearly as feature rich as Cobian? (and free!)

laze.

[just_laze]

Hey,

I thought I'd return and update you all and anyone who searches for 'back-up solution' and comes across this thread.

In short, I now use Microsoft's back-up utility - SyncToy - it's really good and I'd recommend it. All I've done is set-up a rule that 'Echoes' (mirrors) drive A to drive B. Works a treat.

laze