Don't you just love it when (PC Sob story)

Comments

Liam_Vegas wrote on 9/1/2005, 1:10 PM
Tom... thanks for the ideas.... for this particular machine I do normally power it off every night so it's definitely not on 24x7 (or anything close to that). Another machine I have here is left on for days at a time... but I don;t have a problem with that one :-)
GlennChan wrote on 9/1/2005, 1:12 PM
To add onto what Tom is saying:

If your drive isn't spinning up due to physical reasons, you may be able to get it to temporarily work.

Try:
Keeping it well-cooled.
Dropping/slamming it (no joke- this does sometimes work).
Freezing it (watch out for condensation).
see http://techrepublic.com.com/5138-1035_11-728960.html

CAVEATS:
If the drive spins up and you hear horrible grinding noises, turn off your computer right away (with the power switch on the back of your computer's power supply, or pull the plug). The heads of the hard drive are physically hitting the platters that store the data.

if the data is really important, DO NOT DO THE ABOVE. You can make your data less recoverable. Just spend the several hundred dollars to get your data recovered.

If your drive is spinning up, then doing the above won't work. You can hear/feel it spin up.

Static can damage your hard drive, so ground yourself (obviously not on sensitive computer components; do it on the exposed metal on teh back of your computer's power supply).
ken c wrote on 9/1/2005, 1:33 PM
same story... one of my 230 gig internal drives was starting to fail a couple of days ago... gave read errors (inpage operation failure)... it's only a couple years old...

so I defragged it, reboot/chkdsked it, most of the data's now still there.. (whew)... so I've been backing it up to DVDs and to an external drive...

that's a hassle, btw... anyone know any good dual layer DVD burners that would be good for backing data up? doing 8-9 gigs per burn would save time, vs doing 4.3 gigs on regular dvds..

of course that won't help with my 20-30 gig source avis from shoots, so those I'll have to backup on new drives...

Anyone have a great backup system they'd care to share? eg how to back up 8+ x 300gigs of data without it being a weeklong job?

thx..

ken
GlennChan wrote on 9/1/2005, 1:51 PM
ken:

Options I can think of:
A- Tape.

How much data each tape can hold depends on how much you want to pay. They go up to over 100GB/tape (compression lets the manufacturer fudge that number; compression won't work for video).

The tape drive itself is quite expensive though...!

B- Archive your project files and keep your video tapes. You can capture off your tapes since the project file contains timecode info.
You need to backup everything not on your original tapes... i.e. music, rendered out files, etc.
A good file folder structure will really help you back things up.

C- Use hard drives. This is super convenient, although kind of pricey (but still cheaper than A I believe).
You can use drive trays, external drives (drives in external enclosures really).

You can also build a file server with a bunch of drives in RAID 5. RAID 5 provides for fault tolerance, so data is safe if one drive dies (the RAID becomes incredibly slow though, until you put in a replacement drive of the exact model).
Use any old computer, P3 800mhz is good enough.
Put Linux (i.e. Debian) or XP Pro on it.
Get a RAID controller card. You don't need a good one.
Get a bunch of hard drives... I suggest the largest ones you can afford because if you run out of space, you're in trouble.

8X500GB drives = $328 X 8
Computer, ideally with gigE and the right PCI slot for RAID controller = $1000?
The computer can be tricky to get right... this is not like a normal computer. Need lots of power, space, cooling, and probably a server-type motherboard.
RAID controller - $430?

3500GB (with fault tolerance) for about $4000.
With ntbackup (comes with xp pro, downloadable for xp home) you can automate an incremental backup to your backup server (set this to happen overnight).
ken c wrote on 9/1/2005, 2:49 PM
Thanks, Glenn.. some interesting options you've outlined there ... I'll find out more about them... sounds helpful, appreciate it.. challenge is keeping all this important data (especially things like source avis and veg's etc) all archived for When (not if) the original hard drive(s) fail...

gets a bit expensive at $1+ per gig with a couple/few thousand gigs of data... DVDs are good .... but slow re backing up... at least with IDE drives or externals, can just copy overnight for 200-300 gigs..
but that too is a hassle re, like you said, folders, keeping track of what's where..

ken
Liam_Vegas wrote on 9/1/2005, 2:58 PM
So I disconnected my working slave IDE drive and connected one of the faulty drives that were copnnected to the Promise IDE port (setup the drive with the correct jumpers to operate in slave).

The startup complains the drive has a hardware error. After booting to Windows the drive is not listed anywhere (unlike before - when it was hooked up to the Promise IDE port). - Odd.

Anyway... I have a copy of Spinrite (which I used on a true failing hard drive about 6 months ago... and it recovered the data).

So I booted up into the Spinrite utiltiy (that works under freeDOS) and it can see the drive... although it seems to be missing some of the "setup" data about the drive. In comparison to the working Maxtor drive (C) which has lots of additional drive information (SMART DATA)... Spinrite does see this drive correctly in terms of basic drive data (capacity / cylinders / heads/ sectors)

I now have spinrite working on analyzing the data on the drive. Spinrite "sees" the data just fine (I have it running it's level 2 recovery mode) and so far it is not reporting ANY data errors.

That's the frustrating thing here... it really does look like the data is all there... and nothing is actually wrong with it. So why then does the O/S not actually see the drive in a way it can be mounted?

EDIT: so a little more on this. Spinrite sees the NTFS volume/partition... but under "Physical Drive Details" all the data is listed as "unknown"... now is that weird or what?
kentwolf wrote on 9/1/2005, 5:35 PM
>>Are they *logical* drives?
>>No... This is in Device manager - this lists physical devices not
>>logical drives.

Now that I am at home, actually I meant "Dynamic" drives; not "logical."

Sorry.
GlennChan wrote on 9/1/2005, 5:59 PM
Ken, another option for backup is using a bunch of hard drives.

Look for the rebate deals on them. You can check hot deals sites for your country.

For US, I'm hearing about nice deals like 160GB for $50.

A- If you want, you can dig around your case a lot manually installing the drives. After a while this would get annoying.

B- Buy drive trays off newegg (US only). You just swap hard drives around.
They're around $20 each?

C- You can buy an external enclosure and run it semi-naked, which makes things slightly cheaper at the hassle of having to install/uninstall drives. This can be a little more effort than B
Less hassle than A.
Be semi-naked I mean you don't use the screws, so you can swap drives in and out of the enclosure faster.

Prolific chipsets on some external enclosures (i.e. somy Bytecc) need firmware flashing, so you may want to avoid them.

craftech wrote on 9/1/2005, 7:34 PM
Liam,
The onboard or card controllers that are capable of Raid are SCSI devices and as a result they have their own bios that loads after the main one loads. In order for the computer to recognize the drives the boot order has to be correct so the main bios has to load the SCSI device (controller) FIRST. Thus SCSI, A, C. Likewise if you want to boot from a floppy disk with such a setup you must enter the bios and change that parameter to A,C, SCSI or you can't boot from a floppy disk.
The drives have to be connected to the mainboard connectors designated for the controller. If there are four IDE connections often only two will be controlled by the onboard SCSI controller and they have to be connected with 80 conductor cables with the blue end plugged into the proper IDE terminals, usually IDE 3 and 4.
The main bios will show NO IDE DRIVES, but after the first or sometimes second boot screen a SCSI device boot screen for the controller will appear which you can access with designated keystrokes like Ctrl H or something like that.
Then the C drive (primary boot drive ) will load and when Windows loads it will show the SCSI controlled IDE drives properly IF the proper windows drivers were installed. Anything connected to the other two terminals will show up no matter what because they are not controlled by the SCSI - IDE controller. That is how Highpoint works and that is how Promise works as well.

John
Liam_Vegas wrote on 9/1/2005, 8:27 PM
Thanks for the additional description. Unfortuntaley... it doesn't help in this situation. The drives are simply inaccessible - although Spinrite (check earlier post) reports all the data as being intact.

Very frustrating... as it seems the data is just sitting there... but for some reason the drive just cannot be seen at the Windows O/S level.
Catwell wrote on 9/1/2005, 8:52 PM
I have the same the MB as you. I think the promise controller writes special data to the drive which makes them unreadable through any other systems. A couple of weeks ago I had a problem seeing my RAID 0 drive (2 WD 120s). I rebooted the computer and the drive came up. I managed to copy the data to another drive and then I ran Spinrite on the RAID 0 drive. It hung at 57.95%. I removed the two drives and put them in as secondary on the IDE controller. It could not see the drives. I then put one of the drives on the promise controller by itself and I was able to format it and Spinrite found no problem. The second drive could not be seen by anything. Not the promise controller, WIN2K , or Spinrite. I trashed that drive and bought two more. I am running everything as stand alone. No more raid.

Anyway, spinrite has trouble seeing the information on the drives on the promise controller. It can however check the data and surface.

A couple of suggestions that I am sure you have checked. When you are moving the drives to another machine be sure that the jumper for Master, Slave, C/S is correct. This has caught me more than once. Do have additional drives on the promise controller that are working? I ask because the controller can handle 4 drives and you have been talking about two of them. Can you find another machine with a promise controller as you have in your machine? I suspect that your controller has failed but the setup the the controller has written to your hard drives prevents them from working in a standard IDE port. Have you tried putting a different drive on the promise controller? If it works then you need to concentrate on the drives themselves.

Charlie
kentwolf wrote on 9/1/2005, 9:10 PM
Billae :

>>I had a similar situation and used it as an excuse to buy a new system.

Lemme guess...you're married? Right? :)
Edin1 wrote on 9/2/2005, 2:06 AM
So, it doesn' t seem to be a diode.
It is very likely that the logic circuitry has been affected by the surge, which explains the working/not-working, reading/not-reading scenario. I don't know if rebuilding the MBT (Master Boot Table) will help much, but it's worth a try!
One thing is for sure; you will need to send both drives for replacement! If none of the safe methods (the ones that don't destroy your data) work, then you should try everything except opening drives and replacing the platters (you need to have a Cleanroom for that! Trust me, I worked there, and I tried to replace platters at home, only to find that all my data was lost, anddrvie couldn't work because of contamination.).
Liam_Vegas wrote on 9/2/2005, 9:02 AM
A couple of suggestions that I am sure you have checked. When you are moving the drives to another machine be sure that the jumper for Master, Slave, C/S is correct. This has caught me more than once.

Yep.... very careful with that stuff.

Do have additional drives on the promise controller that are working? I ask because the controller can handle 4 drives and you have been talking about two of them.

No... Just two on the promise controller - and two on the standard IDE.

Can you find another machine with a promise controller as you have in your machine? I suspect that your controller has failed but the setup the the controller has written to your hard drives prevents them from working in a standard IDE port.

I don't hve anbother machine handy to do such a test. Also... rememebr... these drives were configured to run in IDE mode... not in RAID mode. In that case would the RAID controller write to the drives in this "distructive/can't use it in any other PC mode"?

Have you tried putting a different drive on the promise controller? If it works then you need to concentrate on the drives themselves.

I have not. I'll maybe try that later today. Just got to pull out a drive that I can bare to lose.

Thanks for the response.
Liam_Vegas wrote on 9/2/2005, 9:03 AM
I don't know if rebuilding the MBT (Master Boot Table) will help much, but it's worth a try!

Do you have a suggestion for how I can do this? Bearing in mind the drives are all but inaccessible at the O/S level?
RafalK wrote on 9/2/2005, 12:57 PM
You didn't by any chance have a Dell, did you?
Liam_Vegas wrote on 9/2/2005, 1:18 PM
You didn't by any chance have a Dell, did you?

Nope.
craftech wrote on 9/2/2005, 1:24 PM
Liam,
Did you ever try the utilities from Western Digital I linked above?

John
Liam_Vegas wrote on 9/2/2005, 2:11 PM
Yes... I did try those utilties. As the drive is basically inaccessible from the O/S - the utilties cannot see them (I tried). That was after I connected the drives direct to the "standard" IDE controller. Maybe I need to connect them back onto the Promise IDE controller (as at least they were visible to "Device Manager" if nothing else).
craftech wrote on 9/2/2005, 8:58 PM
As the drive is basically inaccessible from the O/S - the utilties cannot see them (I tried).
========
Wrong version. I always use the DOS version installed on a floppy and change the boot order in the bios to boot from the floppy which loads the WD utilities and diagnostics. Works very well.

John
Liam_Vegas wrote on 9/2/2005, 10:08 PM
OK... I'll check those out again.... I didn't see that version (must be blind I guess). Thanks
Liam_Vegas wrote on 9/3/2005, 1:20 PM
Tried the DOS utility and it fails on bootup with the following error

"Can't load BDOS kernal file: a:\IBMDOS.com"

so... I guess that puts an end to that one then. Oh well... worth a try,

EDIT

Re-created the DOS boot / Diag disk... and re-run it... got a little further )no idea why the other disk fialed). I got a different error this time

COMMAND ERROR: ERROR/STATUS CODE: 0132

Hmm... wonder what that means....