HDD Fail - What to do?

[Ichan]

It does a no change write in raw mode, so it don't need or change the data.
Think of it as a software hammer, what have you to loose.

I am considering to try it, after the part swapping trial. Browsing for it i got some pro and contra, one interesting opinion is here - do you have experience with this one?

I think the controller swap is worth a try. As Pickit said, what have you to lose at this point? If your data is worth the cost of another drive to try and recover, that's what I would do.

Hunting for a working identical drive is my priority now. My work which is incidentally saved on that disk worth a month of my time - viewing from another side, the projected gains of the whole work is about 100 units of new 1TB HDD.

Just tried the SeaTools for Windows from the disk maker Seagate, silly it can not detect the drive...

-ichan

[Gallymimu]

Not picking on you at all Ichan as this has happened to me before too.

BUT it's a good reminder to be a backup nazi.  It's so important to have an automated backup system.  Cloud backup is cheap these days, as are piles of USB or esata hard drives.  It especially sucks to lose paid project work.  Sorry Ichan I feel ya!  I hated that sinking feeling in my stomach when I realized "it's all gone"

We used to do 5 day rolling backups on regular sata hard drives.  Engineering storage requirements are pretty light unless you have large solid works assemblies or big simulation outputs.

[pickit2]

yes I have used it, and it fixed my old laptop drive, which is still working after 3 years.

http://www.dposoft.net/hdd.html

older version
http://www.4shared.com/get/O3O69ht2/hdd_regenerator_171.html

[zuisti]

yes I have used it, and it fixed my old laptop drive, which is still working after 3 years.
http://www.dposoft.net/hdd.html

Thanks!
BTW, the HDD in question is a SAMSUNG HD753LJ (750GB, SATA3, 7200rpm). Be about 4 years old, but it is very quiet, no any abnormal noise, not too warm. Once the PC (Vista Ultimate) stands up (very slow booting), it works normally, but indicates that some system files could not be reached (eg the FontCache-S-1-5-21*. dat). But occasionally (once or twice a week) here is the "blue screen of death." Then reset and it works again.

As I wrote, I want to do a full backup, and then immediately I will swap the HDD.

[Ichan]

Enough of this, i decided to redo the lost works...

Some story, i got the identical drive then do the controller swap - nada.. both did not recognized by the bios. There is on youtube a video telling that controller swap will not work anymore, as the controller save some disk parameter on the eeprom. Swapping back the controller make both back to their original state.

I then bravely swap the platter.... and now i have two dead bodies...

-ichan

[dotm]

 
I tried to swap platters two moths ago.
Cleaned the whole room twice and cleaned all desks with hepa filter vacuum cleaners.
Build myself a little box with cutouts for rubber gloves and a glass front. Cleaned all tools with compressed air from cans and isopropyalcohol, had a facemask and took EXTREME caution.
Result: headcrash.
Never again.

[gan_canny]

With mean times before failure equivalent to commercial jet engines it makes HDD's very reliable. Nevertheless like airplanes two engines capable of keeping things flying is just prudent. Mirrored Raid drives are the equivalent of two engines. Backups are the equivalent of preventative maintenance. Still with millions of HDD's spinning everyday there will be crashes. If there is a crash the next best thing is to help the injured. This is the equivalent of drive recovery. Thanks for the ideas on crash triage. 

[Bobbla]

I'm probably a little late to the party, but what the heck.. maybe someone else will need this later.

HDD:
IF the HDD dies while making a sound, its a mechanical failure and your only hope is professionals. (This is mainly bad, as the sound usually comes from reading heads plowing the platters.)
IF the HDD dies without a sound, its an electrical failure and your only hope is professionals. (This is mainly costly, unless you have access to a clean room. Ever since HDD's became larger then 20-40GB platter replacement outside clean room, is not repair.)
IF the HDD dies while you are not around to observe its death, its a one of the above and your only hope is professionals.

Cables: IF the HDD starts to get a lot of errors and corruptions, usually its in my experience that cables might be at fault. (This happens over time)
Controllers: IF the HDD starts on one controller but not another.. duh! (Sometimes one of those fancy USB/ESATA slot in thingy's are useful.)

When the HDD starts to hang, give error messages, get corrupted files and other odd behavior:
1. Determine the state of HDD. (S.M.A.R.T data will help you.)
2. Determine the state of data cables and HDD Controller.
3. Replace faulty part(s).

When your HDD stop to be cooperative/responsive:
1. Test on another HDD controller with other cables.
2. Let the HDD rest(disconnected) for a day or two. (For some reason this _SOMETIMES_ work..)
3. Replace HDD NOW!!

On RAID and Backup:
- RAID is not the same as backup.
- RAID is convenient, backup is good.
- Backup can contain RAID, RAID does not contain backup.
- Software RAID is superior to hardware RAID.
- Backups _should_ always be physically(geographically) separated.

That being said..
- You can backup to external HDDs.
- You can backup to home server(s)(aka. NAS).
- You can backup to a external storage service (aka. "cloud").

If you only need a TB of storage.. then you should probably go for an external HDD.
If you need several TB of storage.. then you should go for a self build home server.
I have never used any "cloud" services, I prefer to have physical control over my stuff and it seems safer.
I see some people buy ready to use NAS solutions, but I personally find them to be cost ineffective vs. external HDD and self build home server (Depends on needed capacity.).

I recommend you choose something that does not require to much effort on your part.
If you choose external HDD, NAS or self build home server I recommend you look in to installing rsync: http://www.youtube.com/watch?v=3bdRnZmTecE
If you want to build a home server I recommend you look in to zfsguru: zfsguru (easy to use)


If you have any questions, please ask.  

[Ichan]

Build myself a little box with cutouts...

I use an infant incubator enclosure as the clean chamber, wipe and vacuum cleaned. To handle the head i tried to use a small cut of cable sleeve to lift the head from platter as the first attemp, but then switch to the trick in here.

If there is a crash the next best thing is to help the injured.

Case closed, all the passengers considered died, some probably killed during the rescue operation 

If you have any questions, please ask. 

Is that true controller swap do not work anymore ?

As for backing up, i actually bought recently a 1 TB external drive meant to back up that drive as i already feel it had some delay sometimes - but alas, i handed over it to my woman as she need it for re-arranging her laptop drive content and applications  .

So, on a stand alone computer what automatic backup software to use?

-ichan

[Bobbla]

Is that true controller swap do not work anymore ?

I imagine that it depends on the manufacturer, but I am no expert. (Also, I refereed to the controller cards/adapters, not the one that sits on the actual HDD.)

As for backing up, i actually bought recently a 1 TB external drive meant to back up that drive as i already feel it had some delay sometimes - but alas, i handed over it to my woman as she need it for re-arranging her laptop drive content and applications  .

I guess we all learned a life lesson there. "Helping others is not good for karma."

So, on a stand alone computer what automatic backup software to use?

I recommend you choose something that does not require to much effort on your part.
If you choose external HDD ... I recommend you look in to installing rsync: http://www.youtube.com/watch?v=3bdRnZmTecE

Cheers!

[Gallymimu]

So, on a stand alone computer what automatic backup software to use?

-ichan

depends on how you want to backup as there are lots of options.

I prefer the sync approach and like a few programs: ViceVersa, syncback SE or Pro, and one I just started using that seems very promising (and free) is freefilesync.

[Magnox]

I prefer to use sync software for backups too. My favourite is Second Copy. I do an incremental sync that does not remove data from the backup that no longer exists on the source. That way, if I delete something accidentally, and don't notice until after a backup, I can still recover the file. Of course it also means that occasionally I have to 'clean' the backup HD manually, or just wipe it and do a brand new sync but that takes a while!

I have two 3TB HDs in my main PC, one for relatively non-volatile data (like software downloads and document archives) and the other for backup. That's in addition to the boot HD which has OS, applications and general data like 'My Documents' and some working partitions.

Second Copy is scheduled to run a nightly sync of a few working areas onto the backup HD, plus a weekly sync from the main data HD and some other working areas. Included are a few folders from my main server, where I have my wife's laptop redirected to for instance.

On a roughly monthly basis I plug in a couple of external eSata HDs and run a full sync of the internal backup HD, and some other areas like my AV server's music partition. Those external HDs are kept at a friends house for an offsite backup.

In an ideal world I would also back up my Blue Ray movies collection, but since I have about 15 big HD's full, I can't afford to do that. At least all the Western Digital HDs have now died and the remaining Seagates seem more reliable. I can just about tolerate the loss of one of those. That said, I do have one backup of my real favourites, and some hard to find stuff like fan edits of star wars.

So, backups... yeah, I learned about those after losing data. Now I'm a bit paranoid, but that's a good thing when it comes to data.

[Gallymimu]

I prefer to use sync software for backups too. My favourite is Second Copy. I do an incremental sync that does not remove data from the backup that no longer exists on the source. That way, if I delete something accidentally, and don't notice until after a backup, I can still recover the file. Of course it also means that occasionally I have to 'clean' the backup HD manually, or just wipe it and do a brand new sync but that takes a while!

I have two 3TB HDs in my main PC, one for relatively non-volatile data (like software downloads and document archives) and the other for backup. That's in addition to the boot HD which has OS, applications and general data like 'My Documents' and some working partitions.

Second Copy is scheduled to run a nightly sync of a few working areas onto the backup HD, plus a weekly sync from the main data HD and some other working areas. Included are a few folders from my main server, where I have my wife's laptop redirected to for instance.

On a roughly monthly basis I plug in a couple of external eSata HDs and run a full sync of the internal backup HD, and some other areas like my AV server's music partition. Those external HDs are kept at a friends house for an offsite backup.

In an ideal world I would also back up my Blue Ray movies collection, but since I have about 15 big HD's full, I can't afford to do that. At least all the Western Digital HDs have now died and the remaining Seagates seem more reliable. I can just about tolerate the loss of one of those. That said, I do have one backup of my real favourites, and some hard to find stuff like fan edits of star wars.

So, backups... yeah, I learned about those after losing data. Now I'm a bit paranoid, but that's a good thing when it comes to data.

Good man, takes his data seriously!  I'm not as good about the offsite backups as I should be (I occasionally take a hard drive from home to the office and vice versa).  Otherwise I run my systems similarly.  storage drive in the desktop backed up to sever, server keeps a second copy of all data on another drive, and an external drive that isn't always plugged in just in case of a power surge or someone steals the computers.

Note, I don't believe in using RAID, yes it's the popular thing but remember RAID is designed for maximal uptime NOT for data backup.  It really sucks deleting the wrong thing off of a RAID, guess what, it's gone .  Also I've had bad experiences with recovering date from RAIDS, moving raids from one machine to another (if the RAID controller or motherboard dies) and rebuilds haven't always worked as expected.

[Magnox]

RAID is great for big storage areas and keeping your system up when a drive fails, but as you say, RAID has nothing whatsoever to do with data security.

That's a mistake I've seen several companies make, which has often been a rather expensive learning experience for them! In fact, I think the SCSI drive I managed to recover by swapping the HD's PCB was one of those instances.

A long time ago (in a galaxy far, far away) I did think of building a big RAID server for my movies and music. I even got hold of (very cheaply) a 16-port PCIx SATA RAID controller and started the build. In the end though I realised that I would rather risk losing a single drive occasionally, and deal mentally with the loss, than risk losing all 16 drives if the array went belly-up. Plus, plugging a single drive into a slot-in caddy takes a lot less power than running all of them at once.

[Bobbla]

RAID is great for big storage areas and keeping your system up when a drive fails, but as you say, RAID has nothing whatsoever to do with data security.

What is this??? I suspect you mean data integrity, but whatever.. I would like to say that there are many types of RAID, and I personally use something called RAIDZ which is software RAID that comes with the file system ZFS. Why? because RAIDZ is the ZFS equivalent to RAID5 (n+1), meaning one disk will be used for parity, and thus you can loose one disk and still have all your data. Sure you loose some capacity but you don't loose all the all the data when a HDD dies. In addition ZFS does data integrity checks on all the data, I once had some problem with a SATA cable which gave corrupted some data. I then "scrubed" the file system and it fixed the corrupted data. It also have snapshots which can help you in case you actually delete something you shouldn't have. Also, since RAIDZ is software RAID, it means we don't really care to much about the hardware.. a HDD is a HDD, and if it contains ZFS we can use it no matter what HDD controller card is used.

Basically I could sit here all day writing about ZFS.. but someone has done it before: http://en.wikipedia.org/wiki/Zfs

Also, there are other good file systems like BTRFS and HAMMER.. but I like ZFS 

Cheers! 

[Magnox]

Security... integrity... availability... depends who you talk to.

I don't believe that there is any significant difference between software and hardware RAID from a 'not losing your data' viewpoint. Comparing the same RAID level on each of course.

While with software RAID you might not be vulnerable to controller failure (and then trying to source a replacement) you are more vulnerable to software failure possibly having a fit and scrambling the array. I trust hardware more than software and have replaced dives on many server arrays on HP, IBM, Dell etc with hardware RAID without issues, and seen the occasional Windows software RAID become unrecoverable. Admittedly that's Windows, which no-one should trust to do anything vital. I have tried software RAID ages ago on Linux boxes, but still prefer hardware just from a 'feels better' factor.

Software RAID also has a hit on the machine it is running on, which might not be an issue for a dedicated fileserver.

Either hardware or software RAID can suffer complete loss if two drives die at once (unless using RAID 6) or the box gets zapped by lightening (seen it once at a school - network sockets were melted!).

I have seen two drives die nearly together, where the second died within a couple of days of the first... before the owner had replaced the first HD. Oops! They were another lot that didn't have a backup either... double oops! I sold them a new server, with a backup tape drive.

Personally at home, it comes down to power and noise too. Having one big HD spinning is quieter and more efficient than three. And they both still need a backup!

Oh, and Volume Shadow Copy is great as well, but I don't use it myself because I prefer to just sync.

[Bobbla]

I don't believe that there is any significant difference between software and hardware RAID from a 'not losing your data' viewpoint. Comparing the same RAID level on each of course.

Well, believe one can do in church.. Two sentences from the source of all that is true.

While with software RAID you might not be vulnerable to controller failure (and then trying to source a replacement) you are more vulnerable to software failure possibly having a fit and scrambling the array.

What I said was, or at least tried to say was that most hardware RAID solutions locks you to their products when it fails. This is because there are no real standard, no standard no comparability and you get vendor locked. Also, how is it that the software all of a sudden is going to have a "fit" and shouldn't software (abnormal) behavior be connected to hardware?

I trust hardware more than software and have replaced dives on many server arrays on HP, IBM, Dell etc with hardware RAID without issues, and seen the occasional Windows software RAID become unrecoverable. Admittedly that's Windows, which no-one should trust to do anything vital. I have tried software RAID ages ago on Linux boxes, but still prefer hardware just from a 'feels better' factor.

Software RAID also has a hit on the machine it is running on, which might not be an issue for a dedicated fileserver.

I haven't tested that many RAID solutions, but what I can tell you is that ZFS at least have a very little performance impact on the CPU. It does however require some RAM, this is "tunable" but it requires it for among other this cache. That being said, if you activate compress it will need more CPU resources, or if you use deduplication it will require even more RAM. But these futures are optionals.

Either hardware or software RAID can suffer complete loss if two drives die at once (unless using RAID 6) or the box gets zapped by lightening (seen it once at a school - network sockets were melted!).

I have seen two drives die nearly together, where the second died within a couple of days of the first... before the owner had replaced the first HD. Oops! They were another lot that didn't have a backup either... double oops! I sold them a new server, with a backup tape drive.

Actually ZFS supports up to 3 parity disk as far as I know, and then you can start to mirror... And then there the fact that you can add hot spares, which can replaces the downed HDD immediately. As to your lightning, I not arguing that RAID can replace backup.. but I do recommend some sort of UPS or at least some sort of overvoltage protection device. Also, wouldn't the backup tape get zapped as well if the lightning struck? (at least the tapes that were inserted during said lightning strike)

Personally at home, it comes down to power and noise too. Having one big HD spinning is quieter and more efficient than three. And they both still need a backup!

Power and noise you say, today a HDD's WILL require less then 10W, most idle below 5W. So power requirements are not that high. The noise is a question of placement and case choice. And as for your backup, in my case my little server is the backup for what I have on my workstation. Of course all the TB's of family photos and linux distributions I have on the server is not backed up somewhere else. Why? Because backup costs, at least more then setting up a RAID.

herp derp