Recertified WD20EADS with Advanced Format?

[AC2]

Hi All

Had a WD Caviar Green 1.5TB crap out in 7 months, losing most of the data in the process (no SMART warnings).

Replaced under RMA and have been shipped a recertified (aargh!) WD20EADS (WD Caviar Green 2TB).

Now as per WD's site WD20EARS is Advanced Format while WD20EADS is not, but the sticker on this baby clearly says it's an Advanced Format drive and talks about using WD Align/ jumpers etc if one is using XP.

So how can I confirm if this ruddy thing is really Advanced Format or not?

Thanks!
AC2

 

[RebateMonger]

It sounds like somebody made a mistake with a sticker. Software like HTTune should give the model number directly from the disk.

 

[AC2]

Have run an extended test and zeroising test via the WD Data Lifeguard Diagnostic tool. The tool consistently reports the model no WD20EADS.

The same serial no + model no mentioned on the drive sticker is also mentioned:
a. In the RMA fulfillment email
b. In the warranty lookup for the given serial no on wdc.com

So I don't think there is any mistake on the model no...

 

[sub.mesa]

That is odd; i thought the EADS was the last 2TB drive that still has 512 byte sectors. Perhaps HDD makers are worried about BER on their high capacity drives and thus give you 4K sectors to cause them less RMA'ed parts.

I agree you should test this theory somehow. How? Well you can do random write with CrystalDiskMark; one time on an aligned partition the other time on misaligned partition. the default Win7 partitions are aligned. Use GParted/Ubuntu Linux/WinXP to create misaligned partition and test random IOps on there.

Especially the random write should be at least 50% lower than the aligned partition, if it is a 4K sector drive using 512-byte emulation, like all the Advanced Format drives out currently.

 

[AC2]

I agree you should test this theory somehow. How? Well you can do random write with CrystalDiskMark; one time on an aligned partition the other time on misaligned partition. the default Win7 partitions are aligned. Use GParted/Ubuntu Linux/WinXP to create misaligned partition and test random IOps on there.

Especially the random write should be at least 50% lower than the aligned partition, if it is a 4K sector drive using 512-byte emulation, like all the Advanced Format drives out currently.

Hmm sound advice but seems like a lot of work! I thought I'd check hdparm first to see if it reports physical sector size of 4K (as some of the new EARS drives do) but it doesn't. Instead it doesn't report the physical size at all so the system works it out to be 512b based on number of sectors and capacity.

But the hdparm output (as below) is consistent with that from other EARS drives, posted on the net, which emulate 512b sectors, i.e. the physical sector size is not reported at all:

sudo hdparm -i /dev/sdb

/dev/sdb:

 Model=WDC WD20EADS-11R6B1                     , FwRev=80.00A80, SerialNo=     WD-WMAVXxxxxx51
 Config={ HardSect NotMFM HdSw>15uSec SpinMotCtl Fixed DTR>5Mbs FmtGapReq }
 RawCHS=16383/16/63, TrkSize=0, [COLOR=Blue][B]SectSize=0[/B][/COLOR], ECCbytes=50
 BuffType=unknown, BuffSize=32767kB, MaxMultSect=16, MultSect=?16?
 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=18446744073321613488
 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120}
 PIO modes:  pio0 pio3 pio4 
 DMA modes:  mdma0 mdma1 mdma2 
 UDMA modes: udma0 udma1 udma2 udma3 udma4 udma5 *udma6 
 AdvancedPM=no WriteCache=enabled
 Drive conforms to: Unspecified:  ATA/ATAPI-1,2,3,4,5,6,7

 * signifies the current active mode

I only need to create one primary partition for the whole drive and this is not the boot drive, so I've taken the lazy way out and assumed it is an Advanced Format drive and set up the partition accordingly. If I'm wrong all I've lost are a few sectors of disk space and some time figuring out how to get it working...

> sudo fdisk -lu /dev/sdb

Disk /dev/sdb: 2000.3 GB, 2000398934016 bytes
255 heads, 63 sectors/track, 243201 cylinders, total 3907029168 sectors
Units = sectors of 1 * 512 = 512 bytes
Disk identifier: 0x0004c732

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1              64  3907029160  1953514548+  83  Linux


> sudo mkfs -t ext3 -c -v /dev/sdb1
mke2fs 1.40.8 (13-Mar-2008)
Filesystem label=
OS type: Linux
Block size=4096 (log=2)
Fragment size=4096 (log=2)
122101760 inodes, 488378637 blocks
24418931 blocks (5.00%) reserved for the super user
First data block=0
Maximum filesystem blocks=0
14905 block groups
32768 blocks per group, 32768 fragments per group
8192 inodes per group
Superblock backups stored on blocks: 
    32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208, 
    4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968, 
    102400000, 214990848

Running command: badblocks -b 4096 -X -s /dev/sdb1 488378636

 

[AC2]

Here's a pic of the drive...

 

[AC2]

Done partitioning and formatting as per Advanced Format requirements..

Checking for bad blocks (read-only test): done                                
Writing inode tables: done                            
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done

This filesystem will be automatically checked every 34 mounts or
180 days, whichever comes first.  Use tune2fs -c or -i to override.

Tried copying a 3.7 GB DVD image from main drive (5400 rpm 3Gbps Seagate) to the new drive, done at approx 50 MBps (3.7GB image in around 75 secs).

Copying a large file from /dev/shm proved to be a bit quicker at around 70MBps (960.4MB file in 13 secs). After deleting it from /dev/shm, copying back the same file was done in the same time..

So looks to be OK

 

[RebateMonger]

Very weird.

There's a photos of the sticker on the WD20EADS at Newegg that can be zoomed:
http://www.newegg.com/Product/Produc...4&Tpk=wd20eads

NO "Advanced Format" warning.

Perhaps WDC has decided to offer only a single disk as a 2 TB "Recertified" disk, and that's the 4K Sector version. They send you a 2 TB disk with the same model number, except it's a 4K disk.

Edit:
Here's a quote from the NICX.COM forum, discussing the WD20EADS:
http://forums.ncix.com/forums/index...103&product_id=36130&msgcount=3&overclockid=0
"I know they originally all had 512k but I read that WD changed the newer supply to 4k sectors, keeping the same model # but just a different revision/firmware, that's why I listed that full model # of WD20EADS-00R6B0"

But WDC's current web site shows the EADS as NOT being Advanced Format. For instance:
http://support.wdc.com/product/kb.asp?groupid=608&lang=en

 

[Russwinters]

Knowing WD, I will bet that it IS an advanced format drive.



I could go into a long rant about what drive manufacturers actually do when they refurb drives, but that would just terrify you more.


Buy a new drive for you important data, better yet, buy two and keep a back up.


Use this drive for extra.

 

[RebateMonger]

I could go into a long rant about what drive manufacturers actually do when they refurb drives, but that would just terrify you more.

I would like to hear some details if you can share them.

I've often wondered what goes into "Recertified" disks. It'd be one thing if these are disks that have failed interal QA and are rebuilt. But if they are actual RETURNED disks, I can't imagine it'd be worthwhile for the maker to re-use platters or heads or bearings, since they'd have no idea what the damage might be to those components. That'd leave:
1) The housing
2) The motor
3) The PCB

 

[Old Hippie]

I would like to hear some details if you can share them.

I second that emotion.

You have all your experience just waiting to educate the masses, how 'bout sharing Russ?

If I can't appeal to your sense of humanity, how's a bribe sound?

 

[tweakboy]

Larger drives will have higher chance of dieing. More platters more head work and seeking.

look at my baby still going strong

 

[AC2]

I could go into a long rant about what drive manufacturers actually do when they refurb drives, but that would just terrify you more.


Buy a new drive for you important data, better yet, buy two and keep a back up.


Use this drive for extra.

No worries I'm terrified enough already and had the exact same thought re using this as an extra. Problem is that my existing case is pretty rotten, once screwed in place the rear ends of the hard drives overlap the motherboard and ventilation is poor.

So I can't put in another drive w/o stacking it in the slot next to a current drive (bad idea with the ventilation) or having it intersect with the edge of the graphics card or important headers (including SATA!!) on the motherboard.

Larger drives will have higher chance of dieing. More platters more head work and seeking.

look at my baby still going strong

Well I think I had read a piece of analysis by Google re hard drives that concluded that SMART attributes are only useful in predicting around half of hard drive failures so...

But re your first point there seems to be a perverse trend underway with reduced warranties (5yrs - 3 yrs) and much lower (or unpublished like WD) MTBF figures for 'consumer' drives as opposed to the more expensive 'enterprise' drives.

And a lot of people are saying that the high capacity drives are less reliable. If you see the feedback on newegg.com there are just too many poor feedbacks re high-capacity consumer grade drives.

So what gives? Anand seems to be completely enamoured of SSDs and I haven't seen a single meaningful long term test re these new drives, even though they're constantly suggesting SSD + HDD (typically a high capacity WD Caviar Green!!) on the system builds...

 

[Russwinters]

To answer a few questions about recertification; drive manufacturers have internal hardware/software complex that they use to plug into failed drives, and make adjustments to the drive to bring it back to life.

Drives actually have fairly advanced firmware, and within the micro program code of the firmware is the ability to facilitate "self-scan" where the drive can essentially perform the factory scan on itself again once issued the proper commands. The problem is the commands are not universal and are not outlined in the ATA specification. The commands are completely internal for the manufacturers use so in order to discover them you must have an acute knowledge in AT programming (SC interface helps to especially for WD)

You must also have access to a ATA bus analyzer and understand how to read the output data.

This is something even I cannot do yet.


Anyways, to continue about recertification.

Basically, what manufacturers will do is even if there is a mechanical problem with the drive (I.E. one of the heads or surface of one of the platters has died) instead of throwing it out, the will basically modify the firmware to exlude that head or platter from existence and recalculate the defect lists (basically clear the Grown Defects list, G list, and write all defects to the P-list, the Primary Defects list. Once they have done that they will clear all log files and clear SMART.)

Whats happens is that you have a drive that was...just for example a 3 platter, 6 Head, 1TB disk.

After one of the heads dies, they will modify it to become a 750GB Re-certified drive by disabling one of the heads, and only using a portion of the 5th heads data area or something similar.


Now sometime with re-certified drives there are just firmware failures caused by the programmers error, basically something that would need a firmware update.

Instead of issuing a firmware update to fix it for drive that have not failed yet they will just release all new drives with the firmware update, and let all the other drives die and then replace them with re-certified drives.

The manufacturers can fix firmware problems within a few minutes, so it is not a big problem for them.





On another subject. The higher capacity drives ALWAYS have a higher overall failure rate. This is because the highest capacity drives are always pushing the limits of the manufacturers technology at the time. Because hard drives are mechanical devices they are always liable to fail, not if, when.

The more platters, and the more heads a drive have increases the chance of failure.

I personally refuse to use a drive that has more then 2 platters, 4 heads. Right now that would be a 1TB drive with the newest platter technology, like the 1TB 64MB cache FAEX. Those are 500GB platter design, or 250GB per side.

The newest drives like this one you have is the 667GB platter design.

I can't wait to get my hands on a single platter 640GB w/ 64cache...they should be around soon.

 

[tonyptony]

Russ, is the Samsung F3 1TB a two platter job?

What about the WD EADS 1TB?

 

[FishAk]

Concerning platter count, I don't know if this is 100 percent accurate, but...

I also don't understand what the n stands for, as in (2/n)

 

[Russwinters]

I believe there ARE samsung 1TB that are 2 platters, but the exact models I would need to look up.


The website linked above seems pretty nifty, and fairly accurate.


the "n" is the number of heads, which is usually going to be double the number of platters as platter can support 2 heads (one for each side).

The heads are what read and write the data.

 

[tonyptony]

Neat site FishAk. According to that the 1TB F3 is two platters while the 1TB Green EADS is three.

Hmm, I take that back. The particular model used in the new Essential (-112B2) is not listed.

 

[RebateMonger]

Basically, what manufacturers will do is even if there is a mechanical problem with the drive (I.E. one of the heads or surface of one of the platters has died) instead of throwing it out, the will basically modify the firmware to exlude that head or platter from existence and recalculate the defect lists (basically clear the Grown Defects list, G list, and write all defects to the P-list, the Primary Defects list. Once they have done that they will clear all log files and clear SMART.)

That seems awfully risky. If a head has impacted a platter or if there's a plating problem, that means there could be a loose piece of plating inside the housing, which could cause more damage. Or more pieces could be generated because a head is still running over the damaged platter and might hit the damaged area again.

It sure seems risky to ship a drive back out the door without replacing the heads and platters and cleaning out the insides to ensure there's no particles floating around. Customers are REALLY going to be bummed if they have another early failure with their replacement disk. Folks can understand one failure (which perhaps they caused), but a failed replacement disk is likely to make the customer stop buying that brand and letting all their friends know it, too.

 

[Old Hippie]

That seems awfully risky. If a head has impacted a platter or if there's a plating problem, that means there could be a loose piece of plating inside the housing, which could cause more damage. Or more pieces could be generated because a head is still running over the damaged platter and might hit the damaged area again.

Same thing I was thinkin' and I had no idea that a refurb drive could be "modified" to this extent and still be considered "fit" as a replacement.

Thanks Russ, your insight is invaluable. :thumbsup:

 

[AC2]

Russ, thanks for all the info and especially this piece of advice...

I personally refuse to use a drive that has more then 2 platters, 4 heads. Right now that would be a 1TB drive with the newest platter technology, like the 1TB 64MB cache FAEX. Those are 500GB platter design, or 250GB per side.

I'll upgrade the case and get a new 1TB drive, Hitachi is looking the best to me right now... So I'll have one each of Seagate, WD and Hitachi in the case .

Thinking of using the recertified 2TB WD drive a bit like an offline backup by setting 'noauto' in fstab. Any other way of conveniently enabling/ disabling this drive from being used even while leaving it connected in the case?

I guess the only thing I should be happy about is that because I've got a recertified 2TB drive unlikely that it had originally suffered from physical head/ platter problems that were hidden via firware changes in the defect list etc... The extended scan using WD Data LifeGuard Diagnostic and badblocks in Ubuntu also haven't reported any bad sectors and have reported 2TB of usable space...

 

[Russwinters]

That seems awfully risky. If a head has impacted a platter or if there's a plating problem, that means there could be a loose piece of plating inside the housing, which could cause more damage. Or more pieces could be generated because a head is still running over the damaged platter and might hit the damaged area again.

It sure seems risky to ship a drive back out the door without replacing the heads and platters and cleaning out the insides to ensure there's no particles floating around. Customers are REALLY going to be bummed if they have another early failure with their replacement disk. Folks can understand one failure (which perhaps they caused), but a failed replacement disk is likely to make the customer stop buying that brand and letting all their friends know it, too.

*Most* of the time the heads fail there is no notable physical damage to the discs at all. It is fairly easy to tell even without opening a drive if the heads are making contact with the platters.


Regards,