phew...what a weekend..>.<

[bad_brain]

it started on Friday, I booted my main system in the morning and was surprised because it took damn long (almost 10 minutes), and when finally up the "MS Security Essentials" thingy refused to start with a memory error. after downloading the latest updates it worked again. alright, all fine so far.

on Saturday I was on Skype and noticed some strange lags and video freezes, the Skype process also used about 25% CPU resources. after signing off from Skype: all fine again.
so I already put "new OS setup soon" on my list, but I had/have a lot of work queued right now, so it was something for "in 1-2 weeks".

Sunday: endless booting time again in the morning (about 30 minutes), even more .exe files threw memory errors and refused to start, this time (besides the "Security Essentials") also Process Explorer, which seemed fishy...so I assumed it's a virus (installed some shady software 2 weeks ago) and prepared for a new install. the backups already were "fun" because the transfer rate was sloooooow....20 minutes for transferring the 20MB folder with important textfiles for example.
alrighty, when done I started a new install, went fine, but when I was about to restart after the 2nd wave of updates were downloaded: same again, endless boot times (actually it simply stuck in the "Windows 7 is being started" screen). AHHHH!
I had installed drivers and some other stuff at the same time as the updates, so I thought "oh well, maybe that messed it up", so I started another install...this time it took almost 1 hour to even reach 1% for extracting the install files.
so I started to get really convinced it's a faulty HDD and tried a new install, this time on a spare WD Raptor drive I had lying around...and: all fine...AND super-fast again.

right now I am transferring the last files from the RAID to an external drive ("23 hours left" ), when done I can finally disassemble the RAID and see which HDD is actually faulty...

[ayu]

haha we all have that happen once in a while, sucks man xD
Good that you managed to make backups though

[lilrofl]

I hate the 20+ hour copy, what a pain.

Out of curiosity, how long have you had your RAID set up, and what RAID are you using? We are theorizing about RAID failure in school and it would be interesting if I had a real world number to compare to the mathematical probability.

[ph0bYx]

Good bless backups

I've recently had a little job of extracting some files from a faulty HDD. Took me a day to figure out how to do it (done with test-disk) and a couple of days to extract the data. It literally took 5 minutes to change the directory.

[bad_brain]

@lilrofl
it was a RAID0 with 2x500GB Seagates, been running well about hmm...more than 3 years. I have to add the array went through some heavy stress tests already because every couple of months I do secure wipes of the drives with 35 passes....which means almost a whole day of permanent heavy write activity every time.

I have just finished analyzing one drive (I disassembled the array, which had some annoying side effects because the MBR was on the RAID, and the WD Raptor I've done the new Win7 install on still had LILO on it, but that's another story). I used the nifty little tool from Seagate for it (needs to be on a bootable medium), gonna attach it to this post.

so, here's the result of the first quick test:

both former RAID drives obviously damaged.

the long test (took about 2 hours) returned this:

6 bad blocks.

will analyze the 2nd drive most likely tomorrow (if I have time, have some work to catch up with), and I am curious if it also has 6 bad blocks....because that would mean (at least that's my theory) that something went wrong during a specific write activity and because of the striping the error was caused on both disks at the same time...

after the repair (well, which means "mark the blocks as bad"):

short test passed....we will see how that turns out in the long run.

@ph0
you might find the tool useful too for such purposes, it works for drives of all brands (the only Seagate-only functions are the possibility to check for firmware updates and to create error IDs that can be used in warranty cases)...
and yes, luckily I have my REALLY important files backed up even on 2 drives, but it would have still sucked if I would have lost some of the not-so-vital ones like my mp3s (got them backed up too now, yay!) or game saves...^^

[lilrofl]

ah, the seagate drives kind of screw the math up if they were using the 7200.11 or ES2 firmware because that firmware experienced a higher then normal failure rate. With the 7200.12 firmware it would be about 6% of array failure. So maybe it does has more to do with the number of read writes... the thought of losing all my save games though makes me want to do a backup of them very soon lol

[bad_brain]

oh...interesting...didn't knew about those firmware issues...
ordered a SSD today which I will use for a new Debian install on my main box (the one I had was on the WD Raptor I had to use for the new Win7 one). will do an analysis and fix (if possible) of the 2nd Seagate in the night, and then re-build the RAID, might go for a RAID1 this time because it will be a 100% storage one anyway (and not holding the OS like before, so the speed gain of a RAID0 would be pretty pointless anyway).

[bad_brain]

finally analyzed the 2nd HDD, and it turned out that one was the REALLY faulty one...bad blocks that couldn't even be repaired anymore. so I updated the firmware of the 1st one and use that one as temporary storage now.
the OS itself is now running from a SSD (Samsung 840 Pro), and Wheezy is also installed again to the WD Raptor (before it was still Squeeze).
now I just will have to remove the broken one and I am finally all set again...

in the end it was actually even good it happened, I was thinking about getting a SSD since a while already, but never had a real reason to buy a new HDD (and wow, I love the speed gain!). I also had the opportunity to practice manual boot sector setups in both GRUB and Windows, and the Windows one is MUCH more annoying to set up...especially because of the lousy documentation. "fixmbr/fixboot and done?" as they all say? you wish, my HDD wasn't even recognized as Windows install first, so there was not even the option to pick it as "to be fixed".