Failed RAID :l

[blasterboomer]

Well, after 3 and a half years of 24/7 operation, my RAID 5 array has failed due to 2 drives crapping out in a similar way. Ran a 25 pass zero fill and it writes at 6.30MB/s. its done 5 passes and the reallocated sectors havnt increased, these drives have been reliable. They have run 2 virtual machines, a minecraft server and a website for 3 years 24/7 and didnt fail until now, planning on upgrading to 4x 2tb seagates. 3 and a half years is good for 24/7 use of a consumer HDD i think so 10/10 for WD reliability.

[blasterboomer]

Re: Failed RAID :l

on my first post the 2nd image is "HDD 1" and 3rd image is "HDD 2" heres the error scan results from HDD 2 (HDD 1 is still scanning due to bad sectors). I will post the images from the HDD 1 error scan when it completes, assuming the drive doesnt fail by then

EDIT: HDD 1's error scan stopped at 106GB (hdtune bug?), added image 1.. and 2.. and read/write benchmarks

[stj]

Re: Failed RAID :l

what series drive is it? save me searching that model number!

[eccerr0r]

Re: Failed RAID :l

Looks like a blue drive?

Everyone's going to have hard drives of all manufacturers die on them... Then again I ran my RAID5 4x120GB mixture of seagates/maxtors for ~5 years 24/7 (two of the drives hit 7+ years POH since they were repurposed and wrapped around their POH timers.)

And I don't know... my RAID5's are a mixture of brands whenever possible... currently using a mixture of Hitachis and WDs... Of which I already had a Hitachi fail early on.

[fzabkar]

Re: Failed RAID :l

planning on upgrading to 4x 2tb seagates.

Seagate have been experiencing serious quality problems in recent years. In fact they are facing a class action ATM.

[fzabkar]

Re: Failed RAID :l

It's strange how the two drives have exactly the same number of reallocated sectors (661).

Another thing I notice is that the normalised value of the Power On Hours attribute appears to reflect the life in months.

For example, one drive has a POH count of 56 (= 100 - 44). The hours count of 32377 corresponds to 44.35 months. The other drive has recorded 33585 hours and a value of 54 (= 100 - 46). Once again the hours count corresponds to 46.01 months.

So it would seem that each additional month of life decrements the normalised value by 1. In other words, the drive is rated for 100 months.

[blasterboomer]

Re: Failed RAID :l

The drives are caviar SE, not WD blue. Wonder what causes the very low write speed though

[blasterboomer]

Re: Failed RAID :l

Seagate have been experiencing serious quality problems in recent years. In fact they are facing a class action ATM.

are ST100DM001 any good?

[fzabkar]

Re: Failed RAID :l

are ST100DM001 any good?

The "DM" drives are members of the same unreliable Grenada family. Same physical components, same firmware architecture, just different numbers of platters and heads (1TB per platter).

[blasterboomer]

Re: Failed RAID :l

in my server i had a ST2000DM001 and it worked fine 24/7 for about 1 and a half years, when moving it to another server the drive is detected with correct model and serial, model and capacity but it shows as not initialized... when running an error scan in hdtune every block red (bad) and there is no data in the health tab, the pcb also gets warm in about 5 secs around the ram/cache chip but not too hot that i cant keep my hand on it the drive also does not spin up. any hope of reviving it? i really need the data from it, could i swap the pcb from another drive with the same model. The fact that the drive identifies itself correctly with its serial number and capacity gives me some hope

[fzabkar]

Re: Failed RAID :l

When swapping PCBs between these drives, you need to transfer the 8-pin serial flash memory chip, or its contents, from patient to donor. This chip contains unique, drive specific "adaptive" information which is generated during the calibration procedure at manufacturing time.

That said, is it possible that the drive is powering up in standby mode (PUIS)? If so, you can disable PUIS using a tool such as HDAT2. Use the "w" switch to wake up the drive, ie ...

HDAT2 /w

[eccerr0r]

Re: Failed RAID :l

Ugh I have a 2T Seagate spare that I have yet to take out of the package and use, it's been sitting there for months. I should beat the heck out of it before the warranty expires...

[fzabkar]

Re: Failed RAID :l

Ugh I have a 2T Seagate spare that I have yet to take out of the package and use, it's been sitting there for months. I should beat the heck out of it before the warranty expires...

Warranty replacement drives are usually refurbished units. Not to be trusted, IMHO.

[eccerr0r]

Re: Failed RAID :l

Better than nothing however...

[Stefan Payne]

Re: Failed RAID :l

As strange as it sounds but it's best to use different drives for a RAID array because the same drives can die in the same way due to age/wear and tear...

[hikaruichijo]

Re: Failed RAID :l

From my experience when a drive reaches the maximun relocatable sectors it won't mark more in the SMART no matter what you do. So I think the two drives have the same sectors marked because of this.
You could try HDAT, it another good tool.
But that drives have to much relocated sectors to be reliable and fast.
Relocated sectors tend to increase and they increase seek time a lot because of the increase in head movements.
Seagate had the 7200.11 series that failed a lot, and WD had another series very prone to failures and Hitachi....
That happens from time to time.

[RJARRRPCGP]

Re: Failed RAID :l

in my server i had a ST2000DM001 and it worked fine 24/7 for about 1 and a half years, when moving it to another server the drive is detected with correct model and serial, model and capacity but it shows as not initialized... when running an error scan in hdtune every block red (bad) and there is no data in the health tab, the pcb also gets warm in about 5 secs around the ram/cache chip but not too hot that i cant keep my hand on it the drive also does not spin up. any hope of reviving it? i really need the data from it, could i swap the pcb from another drive with the same model. The fact that the drive identifies itself correctly with its serial number and capacity gives me some hope

Sounds like a bad board on the HDD. Or bad soldering on the HDD board.

[fzabkar]

Re: Failed RAID :l

It won't hurt to swap compatible PCBs between ST2000DM001 drives, but this will not give you access to your data. However, if the drive spins up with the donor PCB, this will at least confirm that the original PCB is faulty or that PUIS is enabled on it.

Seagate's drives have a serial diagnostic port. Newer models disable this port, but earlier models provide useful information which can be accessed via HyperTermial, PuTTY, or a similar application. There is lots of information at The HDD Oracle or HDD Guru forums.

[blasterboomer]

Re: Failed RAID :l

I just tried the 2tb seagate in another PC and it worked fine, spins up, had all files, no errors in the SMART data at all. put it back in the other pc and it didnt spin up, didnt show smart data, showed as uninitialised.... any fix for this? the motherboard it works on is an asrock fm2a88x itx+, the one it does not work on is an msi 760gm-p23(fx) with latest bios. Im now going to put the 2tb seagate into a broken WD mycloud i have. lets hope it works

[blasterboomer]

Re: Failed RAID :l

put the wd firmware on the seagate hdd, plugged it into the wdmycloud. drive doesnt spin up.. so its not related to the motherboard

UPDATE: i used hdparm to disable PUIS. The command i did was "sudo hdparm -s 0 dev/sda". The hdd now spins up, after looking at the following link i think this is what may have caused it in the first place but wasnt noticed because the drive stayed in the same pc which the bios supported PUIS