Is my drive toast?

[mubs]

I last checked the health of my drives a few days ago, using Crystal Diskinfo. Smart values were good, with all the 3 drives being healthy. I shut the system down yesterday morning as I was going to be out the whole day. I turned it on again at around 8:30 pm, did some work, and shut it down normally.

I turned it on some 20 minutes ago, and it took an awfully long time to boot. Disk activity light was on solid. After waiting eons, got the log in prompt (it normally comes up pretty quick); after log in, system was unresponsive to mouse actions and the Windows key on the keyboard. Ran Task Manager and then Resource Monitor from there; disk 0 was crazy active. Rebooted the system. Was much more responsive this time, but the disk activity started again. Resource Monitor again showed heavy disk activity on Disk 0. I managed to run Crystal Diskinfo. After much waiting, the screen popped up.

This time, I'm getting a Caution status instead of good on Disk 0. Reallocated sector count and Current Pending Sector count have yellow status in the SMART screen. Shots attached. I guess I better replace the drive asap. Fortunately I had imaged the OS about 10 days ago and done a full back up of the contents a week ago.

This is my first disk failure; I keep my disks cool and have generally replaced them before they start acting up. This drive is 3 - 3.5 years old.

All advice appreciated. Thanks.

 

[LunarMist]

Five reallocated sectors is not the end of the world, but your drive does have almost 10,000 hours on it. I don't know what is normal for those particular model drives, but I have some drives a few reallocated sectors for a long time.

 

[mubs]

Thanks Lunar, but it appears the drive is failing at many spots. I'd rather not take the chance.

I did a surface scan of most of the partitions; on the second one, Windows said it found problems and fixed them. Scan kept "hanging" on the last partition, so I did that by booting into Partition Wizard from a CD. It too kept repeatedly trying to read the same spot over and over again right at the start of the partition.

I aborted that and checked the Windows partition with Partition Wizard. It completed and said no errors were found, but it started off really fast then slowed down to a fifth of the speed and spent a long time at many, many places.

For the immediate situation, I'll stick in an older Hitachi 250 GB as my boot drive and later transfer that to a Samsung 840 250 GB SSD. The rest of the stuff is backed up, so I don't care if the drive goes bye-bye after I've moved Windows off the failing drive.

I bit the bullet and ordered two HGST Deskstar NAS 3.5-Inch 3TB 7200RPM SATA III 64MB Cache Internal Hard Drive Kits (0S03662) at $145 each from the local Amazon. These come in retail boxes, so at least they won't get abused during storage and shipping. I thought of buying 2TB Toshibas, but those are bare drives and seem to have high failure rates.

I'll replace my three drives with these two. Even 3TB is overkill for me, but the second one is going to be a copy of the first so I have a near real-time archive / back-up of my stuff.

I've read that these NAS drives are louder, and one guy said they've made the drive shells thinner now and the drives vibrate. Will have to see how they pan out.

 

[Mercutio]

The NAS drives mostly seem to be fine in my experience. Your current drive will probably even survive well enough to do a normal disk imaging with TrueImage/ImageX/HDDclone as long as you tell them to ignore errors beforehand.

Usually, a drive that's entered a failing state on reallocated sectors will have its disk activity pegged at 100% in Task Manager, but if you want to get a better idea of what's going on, DFT is probably still the way to go. Sometimes third party tools don't fully articulate what is or isn't a fail condition.

 

[Bozo]

I think I wouldn't shut the computer off until you have the new drives in hand and ready to install. It might not boot up.

 

[mubs]

Aaaaaaaaaargh!

When it rains, it pours!

I cloned the Windows partition to the 250GB Hitachi and actually booted up with it. Windows wanted to install drivers for it and reboot. On reboot, the memory fault LED on the motherboard lit up. After powering off and pressing the 4 SIMMs in, that went away, but the VGA error LED is now lit. I am using the GPU built into the Intel 5 series CPU. The DVI connector is screwed in right; in fact everything worked fine before the last reboot demanded by Windows and I didn't touch a thing.

What do I do now? I have a discrete video card from my previous PC; IIRC, it's PCI Express. Should I stick that in and try? If I do that, will the Intel GPU be disabled automagically?

Thanks.

Posting from daughter's laptop. Crummy Windows 8 drives me mad.

 

[mubs]

Well well well! The vide error LED went off on its own, and after a reboot I'm posting from the desktop. Go figure.

Should I sometime soon take everything apart (CPU, RAM), blow compressed air into the sockets and re-seat everything? The CPU heatsink was the Intel retail one with the thermal tape that won't work anymore once I take it off, so I'm leery of messing with it. Advice please?

 

[Bozo]

Old saying "leave sleeping dogs lie"

I've had the same thing happen to me. It's Windows problem. I believe that Windows has a hard time finding where everything is on the new hard drive. After a couple of reboots, it's fine.

 

[mubs]

Old saying "leave sleeping dogs lie"

Amen! Seems even dusting the insides off causes some flakiness.

When I booted off the newly cloned version, some desktop shortcuts disappeared, and Chrome (the browser) lost all its extensions.

A few hours ago I cloned it again to the SSD, and Chrome lost all its extensions again. What gives?

One of the 3TB Hibachi's is supposed to arrive today; it's already > 6pm, but they supposedly deliver till 8.

The DFT tool Merc linked to above will not see the (bad) drive in a USB dock, even though the documentation says it will. Using the Scan option did didly. Since I'm running off the SSD, I removed the spinning OS drive and stuck the bad one in its place. Now DFT sees it. I asked for an extended test. Three blocks came up pretty quick on the progress bar, and nothing after that for about an hour and a half. The disk light on the case blinks very sporadically; it could be the usual I/O to the SSD, though. No way of knowing if the DFT is actually doing something or not. I'll kill it after a few hours if it's not progressing.

 

[Mercutio]

DFT normally runs a single test and progresses sequentially through each drive that has been selected for testing. DFT actually works on all brands of spinning drives and definitely does see externally-attached disks. It also doesn't have a stupid "I can only show one screen full of drives at a time" limitation, unlike Seatools, which is one of the reasons why I'm particularly familiar with it (Seatools, to its credit, runs tests in parallel. That is awfully handy when you need to run eight six and a half hour long extended diagnostics on your brand new 4TB drives).

 

[mubs]

Well, the DFT UI isn't very helpful. I first tried running Scan, but apparently that only checks Smart something that runs for hours. I finally figured out that clicking Utils / Surface Long Test is the option I was wanting. After running 2+ hours, progress bar is at 2 blocks on a 1GB drive. No prob, I'll let it run. Since you recommended it, that's good enough for me.

I checked the warranty on my new HGST 3TB on their website, and it shows 22 months instead of 36. I've contacted the seller, telling him to inform HGST of the sale date and am waiting for an update. Haven't opened the retail package; the serial # was on a sticker outside. I paid extra for expedited delivery and that has come to nought because of this issue. Worse, the second drive is on the way and I'll have to get its warranty issue fixed as well.

 

[Buck]

Yes, DFT on a HGST drive can take some time. This is a 512n sector drive, which will increase the test time. Also, DFT will take control of sector relocation on an HGST drive, which can take a lot of time. Either way, I would replace it and end my use of the drive. We don't know what data is in the relocated sectors. If it contains anything accessed often by the OS, performance will always be miserable. The spare sectors are on the inside track of the drive, which increases data access time. The drive has a sector allocation table, and if that table is maxed out after DFT's work is done, it will fail the drive.

 

[mubs]

Thanks Buck, good to know these details. The drive is out of active service now, and I'm running DFT more out of curiosity. If it is serviceable, it might become one of the 3-5 backup drives I use, or a scratch drive.

I guess it would be prudent to run DFT on the new 3 TB HGST drives before putting them into active use. It will probably run for about 36 hours each? 14+ hours at the half-way mark for this problem 1 GB drive now.

 

[Buck]

Mubs,
I would expect DFT to run much more quickly on these drives, if they are the new versions of their product. Part of this is because of they way 4k clusters are identified and tested. Also, if the drive is error free, the time for a test to complete will be quicker.

 

[mubs]

Thanks Buck, that makes sense.

Meanwhile it's still chugging away on the 1GB - 25 hours and counting, 16 blocks done, guess another 8 to go on the progress bar. I would think if it maxed out on problems it would stop, but it's still going. Now I'm curious as to the end result.

 

[mubs]

60 hours exceeded and still not done. There's still a quarter progress bar's worth of white space at the rightmost end, and it's been in that state for the last 5 hours with the drive access light on the case on solid for the last 60 hours. Is it even doing anything? The UI certainly sucks enormously. Shouldn't the progress bars be evenly sized such that they fill the space evenly instead of needing fractional bars? Crappy programming.

I have two new 3 TB drives to test, and Merc said DFT doesn't multi-task. I'm probably looking at a few more days of this nonsense.

 

[mubs]

DFT Completed. I need some help folks.

For the new 3TB drives, should I run a) DFT or b) just check SMART values + do a long (not quick) format in Windows?

I ask because for the bad drive, DFT now says it is fine, and SMART by Crystal DiskInfo went from caution before the DFT to Bad after. Help please?

-------------------------------------------------------------
         DFT-Win HGST HDD Diagnostic TEST

         Copyright (C) 2012 HGST
-------------------------------------------------------------

      DATE       :  05/16/2015 23:00:49

DRIVE INFORMATION
        Serial Number     : JP2930HQ1NPHNH
        Model             : Hitachi HDS721010CLA332
        Firmware Rev.     : JP4OA39C
        Capacity          : 931GB
        SMART Status      : Healthy

Test Log

        ReadErrStat       : Done
        Result            : Pass
        Date/Time         : 05/13/2015 19:31:42

        SMART ET          : Done
        Result            : Aborted
        Date/Time         : 05/13/2015 19:31:42

        ReadErrStat       : Done
        Result            : Pass
        Date/Time         : 05/13/2015 22:16:44

        SURFACE QT        : Done
        Result            : Aborted
        Date/Time         : 05/13/2015 22:16:44

        ReadErrStat       : Done
        Result            : Pass
        Date/Time         : 05/16/2015 21:39:43

        SURFACE ET        : Done
        Result            : Pass
        Date/Time         : 05/16/2015 21:39:43

So which one to believe, DFT or Crystal DiskInfo's Smart values? ALso note that DFT's very first look at the drive says Smart values are ok (right at the top of the log).

Aother weirdness, my motherboard has an EFI, and for the first time during ownership (2+ years) I saw an AMI Bios screen on cold boot this morning! It asked to hit F1 to enter setup, I did, and it went into the UEFI bios screen. Any explanations??!

 

[ddrueding]

For new drives I do the following: Nothing.

Don't have data without backups. If you do have backups, don't bother with such effort to further reduce such a small risk.

Just my $0.02.

 

[mubs]

Thanks DD; I'll probably do that. Windows popped up a message that a HDD was bad. So I guess the drive is toast. Puzzled though why DFT says it's good.

 

[Mercutio]

When I get a new drive, I run an extended diagnostic test before I put it in service. I'm mostly concerned about shipping damage or the possibility of preexisting bad sectors. The tests take maybe half a day and have saved me from putting DOA drives in arrays a few times.

 

[ddrueding]

Thanks DD; I'll probably do that. Windows popped up a message that a HDD was bad. So I guess the drive is toast. Puzzled though why DFT says it's good.

View attachment 909

I've seen that a few times, but a SMART scan normally shows what it was complaining about.

 

[time]

The SMART scan shows that it needs to reallocate a ton of bad sectors, but has only successfully reallocated one. At least, that's how I'm interpreting it.

The drive IS toast.

 

[mubs]

Wrapping Up!

Thanks all, and time.

But why did the drive manufacturer's utility, DFT, say the drive was good? That too after running for 63 hours on a 1GB drive that was on an Intel SATA 6 Gbps channel. This is what puzzles me and makes me lose confidence in it. Not only SMART, but even the much maligned Windows (7 Ultimate) popped up a window that the drive was heading towards impending doom and I should quickly back up all data and get rid of it. Of course by now the drive was really toast. My first clue that something was wrong was how unresponsive my system was on a cold boot.

For the two new drives, I skipped DFT but checked SMART immediately after initializing them, then partitioned and slow formatted each partition (this is supposed to check for bad sectors) and then again checked SMART. Wth all my data restored, SMART still looks good, and the SSD and two spindles are running at 32, 34 and 33 C respectively. The two Samsung 320 GB drives I pulled out (perfectly good but decided to retire them) always ran a good 4 - 5 C cooler. These Hibachis always seem to run hot.

I am also still foxed by why some of my desktop shortcuts and all of the Chrome browser's extensions disappeared each time I cloned the OS to a new drive.

 

[ddrueding]

Actually had two drives fail on me this morning, might be first of the year.

Original OCZ Vertex 120GB - Fine to undetectable after a reboot.
Brand new 6TB WD Green - DOA, packaging was terrible, click of death on first power-up.