7 replies to this topic

[SchrodingersCat]

I apologize that this is a long read, as I am still at a loss for ideas and thus I am laying out most of my attempts, habits, designs, and thoughts in the hopes that it will allow someone more knowledgeable than I to come to a solution.

 

~~~~~~~~~~~~~

The morning of Tuesday, Oct. 20th had me awake to discover the power was out. My computer normally is online at all hours, at least partially, and thus it was on when the power must have dropped. That being said, I have a Kill-a-watt leading into an APC surge protector, leading into a Corsair RM1000 80plus Gold certified PSU. None of these devices suffered ill fates. I returned from school to find the power restored. I turned the computer back on to encounter the first bad omen, Windows was taking an obnoxiously long time to boot -- over a minute when it normally took a couple seconds, literally. I log in to windows to be greeted with a plethora of "this isn't running", "could not update this", "such and such did not start" etc. errors. Figured it was just some extra nonsense from a bad shutdown until I noticed the shortcut to my main drive was dead. Go to computer>drives. Only two partitions are listed when there should've been 4 (note: 2 of those partitions were on one drive, the affected drive, drive A).
 

An aside. My system runs off of a "large" HDD split into two partitions, one containing media, the other containing games, other programs, and miscellaneous files. This HARD disk drive will hence forth be referred to as drive A including its two partitions for simplicity.

A second drive is a recent SSD purchase to put some random files on as well as big, hefty, fancy games such as GW2 that I want to run at peak speeds. This SSD will hence forth be referred to as drive B. 

A third drive is my original SSD and is hardly big enough to hold windows 7, often even needing to be pruned periodically to make space. Not great I know, but was hoping to buy a replacement for it soon. This Windows only SSD will be referred to as drive C (because that makes sense after all). 

I reboot and hope that resets everything. no dice. So I get into the machine and unplug drive A from power and data, plug it back in, reboot. No dice. I go into the admin tools and pull up storage information.  Drive B and drive C show up, drive A is no-where to be seen. Not even an unformatted partition, corruption errors. again, it's as if the drive wasn't even plugged in despite the contrary. Hoping this might just be an OS error I reboot and go into the BIOS. BIOS doesn't even recognize that the hardware is plugged in. It shows an empty SATA port. I shut down the system and remove the drive for its safety and go to bed.

 

Drive B already had GW2 on it, the next night I had a TLC rally to run and installed teamspeak on drive B, put up with my computer's gimped functionality and had some fun. By Thursday Oct. 22nd, I started considering my options about data recovery which is way out of my budget and where I might be able to cut corners or what options I might not have tried yet. For now I set into the idea of partially rebuilding what I lost and start installing some other core programs and games on to drive B in the hopes that I could at least be mildly ready for HoT the next day. In the middle of steam installing TF2 and Rocket League to drive B it suddenly returns "disk write" errors as if the disk is full. I was sure it had enough space, but I pull up my computer to check its capacity to find that in the middle of installing things to it and with computer operation otherwise nominal, Drive B has now also completely disappeared without a trace! Just like drive A, no recognition from the OS nor the BIOS. And this time there was no power outage to blame it on. As of the time of posting this I have no real explanation as to what truly caused this. At the time it seemed logical that the problem was now likely in the motherboard, rather than the drives, since two drives failing within 2 days of each other in nearly identical ways is just too coincidental. Yet somehow, the oldest drive, also the OS drive, drive C, still functioned. For safety I shut down the computer and it hasn't been turned on again since, over a week later. 

 

Sickening thoughts about the depth of the damage and work lost began to sink in combined with my new job and I'll be honest, I just didn't want to face these facts for a while and let the time go by a bit. The computer talked about this whole time was a recent build put together just this January with almost all brand new, high-end parts to replace a very aged gaming rig I built nearly a decade ago. That rig still works for the most part and I reassembled it to become a testing/troubleshooting machine. After over a week I finally got the old machine running again and after fighting some POST codes after not being used for so long, it was happily booting to BIOS and the freshly burned copy of Ultimate Boot CD (UBCD) I gave it. I plugged in drive A and set to testing. First just seeing if the BIOS of a different motherboard recognized it. No dice. Using a variety of HDD testing tools on UBCD all of them either ended up hanging indefinitely or spitting out errors about not being able to properly read the drive...or both! so no dice. UBCD has a LARGE list of tools to use though and I have hardly exhausted that list. For now though, with everything throwing errors about failing to read the drive, it seems my original hunch that the PCB that runs it is somehow badly damaged.

 

The good news is that the drive does not make the tell-tale "clicking noise of doom". It DOES spin up though which is obvious now that I have it sitting in an area where I can easily feel and listen to it. It seems like it's spinning slower than normal too, but I could just be perceiving it wrong. I have not gotten to test drive B yet. I worry that it has suffered the same ill fate. It is a samsung 840 EVO purchased a few months ago that is still likely under warranty. Sadly, not long after installing it I moved some files to it that exist no where else, which makes claiming a warranty where I lose that data unlikely.

 

The round up, Drive A seems to have a fried PCB, drive B likely does too. Handing them off to a repair/recovery service will cost thousands of US dollars which is significantly out of my budget. I have no idea what caused these events to happen and I feel if I were to leave my main computer online with drive C long enough it would be damaged too. At least that wouldn't hurt much, it's just windows. Losing the other two drives hurts...a lot. For what it is worth, yes, I have a back up, but it's a year old at least. I was getting ready to install a RAID1 array in a month or so to eliminate the need for back ups and got lazy. Too little too late I guess. Along with costly data recovery, until I can pin point the cause of these failures I feel simply replacing the drives could lead to the replacements being destroyed as well. I've so far found no visible damage to the motherboard and aside from drive failures the rig operates fine as if nothing was wrong. I suppose I could test voltages on the PSU to see if maybe it's frying things, but that's about all I can come up with. 

 

Any ideas regarding further troubleshooting of the causes of these failures, or how to restore the damaged hardware, would be greatly appreciated. Any information or field connections anyone has to drive recovery services or employees that might be able to lend me advice or discounted services would also be greatly appreciated.

 

For now. I've spent over an hour writing this and need to go to bed. Thank you all very much for reading that and thanks in advance for any assistance you can offer. 

~~~~~

In hindsight I note that the BIOS chipset had SATA devices set to an IDE standard format as per defaults, and I did not test with the other available settings. While I do not believe this will make a notable difference I will try the other options later and report back.

[Nox]

You really only have two culprits. 

 

A bad SATA or dirty power from the psu.

 

Unfortunately either one could fry a drives PCB.

 

EDIT:

 

Also any form of Raid is not a backup, just another form of redundancy.

 

Might want to rethink how you do your backups.

[Icie Dedpepul]

What Nox said. If drive A went bad, then drive B went bad, it's more than likely your power supply is spiking. The MB could be having voltage control issues, but PSUs blow way more often, then cause all sorts of crappy issues until it's finally replaced. See if the manufacturer of your PSU will take a look at it as a first step.

As for the drive, It could be the drive's PCB is only bad, and the drive itself is ok. I've seen a replacement drive PCB bring a drive back to life. I'm not gonna tell you it's a solution, and it's probably likely the drive is fried, but, it's something to think about.

 

RAID1 is ok, as long as it's a separate unit. Internal RAID setups can suffer the same fates as the other drives if they are part of the chain. Or get a double-the-internal-sized external HD and do automated incremental backups. I have both.

[Faulty Krew]

Just take it to micro center and get it recovered. That is if the data on there is worth more than 100$

[manz]

Also get an AVR/UPS (or at least a surge protector) to protect your unit from power spike cause that normally happens during power outage. 

[SchrodingersCat]

Also get an AVR/UPS (or at least a surge protector) to protect your unit from power spike cause that normally happens during power outage. 

That being said, I have a Kill-a-watt leading into an APC surge protector, leading into a Corsair RM1000 80plus Gold certified PSU.

 

I was already running a supposedly good surge protector. A UPS device was on my shopping list but low priority because I should've still been fine with what i had and I am not rich and can't buy it all at once. But yes, the UPS is now getting bumped up in priority...as soon as I get a machine running again to plug into it that is.

 

 

Just take it to micro center and get it recovered. That is if the data on there is worth more than 100$

I would be willing to spend a few hundred dollars and am hoping I can work out a deal with Microcenter or a local service considering I am not a business, have some basic connections with people, and have done some of the diagnostic work for them already (not that they'll believe me, but I can try). If not maybe at least I can track down an identical PCB in some form for a good price and try it myself. But I am a college student that doesn't even really generate $1k per month before loan payments, current class payments, gasoline, etc. And I don't even pay rent at the moment. So the normal costs of the repair are possibly years off from now at the moment. But I will be holding on to said drives.

 

 

The current running theory is the PSU is faulted somehow, but I haven't gotten to test it yet. If basic measurements with a multimeter don't confirm the damage I might need to get one of those damn PSU testers, but most of them look crappy. 

Regarding the RAID array. I plan on using RAID 1, which is just one drive completely mirrored onto 2+ other drives at a 1:1 ratio. No fancy tricks, just write the same data to multiple, separate drives. That way if one drive is "murdered" the other(s) function as direct backups. All of those drives in the array would have to fail simultaneously to leave no back up possibility. Is that what you guys are saying is bad? I'm not sure I understand why RAID1 would be considered as risky as the others?

[Nox]

 

Regarding the RAID array. I plan on using RAID 1, which is just one drive completely mirrored onto 2+ other drives at a 1:1 ratio. No fancy tricks, just write the same data to multiple, separate drives. That way if one drive is "murdered" the other(s) function as direct backups. All of those drives in the array would have to fail simultaneously to leave no back up possibility. Is that what you guys are saying is bad? I'm not sure I understand why RAID1 would be considered as risky as the others?

 

Its not risky. But its not a back up either. I'm all for that type of configuration, but at the very least you may want to get an external drive to back up to, and/or a crashplan type of service. Up to you of course, and it really depends on what your data is worth to you.

 

I mean you literally just had your computer kill two drives in short order...which just shows that anything can happen and multiple drives can fail at the same time.

 

I mean I have a full multi terabyte ZFS raid at home, and I'm under no illusion that its a "backup".

 

I also have 200+ tb on raids at my job, and although we do disk to disk backups, we also have offsite tapes etc.

 

Raid's are awesome, but they are a redundancy, not a backup. 

 

/Sorry, sysadmin rant is over.

[SchrodingersCat]

Welp. Microcenter contracts a company for data recovery. $800 deposit to even look at a drive, let alone fix it. Gotta find a different tree to bark up I guess.