Matthew J. Harmon: Minneapolis, Minnesota, États-Unis, Earth

Action and Adventure on the High Seas of Information Security

Western Digital Dead Red Drives

Dec 08, 2014, Posted in archive

Archived from Google+

In the future I'll be looking at BackBlaze's drive reviews before purchasing.

My previous media server is reaching the end of its life, 5 years of problem free storage until recently a couple drives started showing some errors and PATA drives, along with my 5yr old CPU socket, have since gone the way of the Dodo. So a great excuse for a new build.

My build was based on a need for reliability, speed, running cool and quiet with low power and of course security (in that order) for my storage. So I decided to step up and build out a RAID array with ZFS on top of LUKS (whole disk) within a Hardened Gentoo build.

Keeping that in mind here are the final build specs:

Motherboard: +ASUS P8H67-M PRO/CSM, Micro ATX form factor with 2 6Gb SATA and 2 3Gb SATA
Processor: Intel Core i7 3770K 3.5GHz
Water Cooler: +Thermaltake Germany TTPro (2x120mm fans + radiator + closed loop)
CPU Memory: 32GB Kingston 1333 (CAS 9-9-9-24)
Boot/Root/Swap: +Samsung USA SATA III 6Gb/s 250GB SSD (supports AES-256 on-board)
Data Storage: 4 x 3TB +Western Digital RED Drives (WD30EFRX)
Case: +Corsair Vengeance C70 (steel + handles + 6 3.5" drive bays + fan mounts everywhere
Power Supply: +Coolermaster Deutschland 650W (only one with 6 SATA on hand at Microcenter)
Backup Power: APC UPS BR700G (450W)

During the overnight memtest burn in (stress test) of the Kingston RAM there were no issues and the system ran cool (< 45C) for the whole test. During my build of Gentoo, the CPU stayed below 40C and power usage didn't go above 110W as reported by my UPS. When idle, the system uses around 77W of power and remains around 31C. The drives remain around 29C. A great testament to the effectiveness of the Thermaltake water cooler and the airflow of the Corsair C70 case.

Now all of this is fine and dandy and it is amazing how fast Gentoo builds. However, I have had two of the Western Digital RED drives die in the past 24 hours. The first one gave out during the LUKS + ZFS build causing all sorts of I/O issues and unrecoverable degradation of the RAID resulting in the ASUS BIOS to report S.M.A.R.T. failure on next boot with the dire message of "backup and replace" (thanks ASUS!) I tested this drive on another system and it reports 860GB of storage available and no S.M.A.R.T. The second drive croaked this morning and now reports a total available capacity of 600,332,565,813,390,450 bytes [600 PB] via smartctl and scsiPrintSasPhy Log Sense Failed [scsi response fails sanity test].

I started an RMA on the first dead drive last night but my confidence in storing anything on these WD RED's is very low right now. I think I need to engage Western Digital to replace this whole batch. This is the first time I've used WD in ages, has anyone else had this happen to them with WD? I know the WD RED drives are intended for small NAS in SOHO environments but is there that much of a difference in quality from the Enterprise grade SATA III drives like the +Samsung USA / +Seagate Barracuda's?


David Ford - Dec 2, 2012

i haven't used any WD RED drives, i can't say :/ my last 3 drives have been OCZ SSD drives thru amazon (2x120GB, and 1x180GB) in separate use. all of them are working dandy.

David Ford - Dec 2, 2012

i do intend to build a nas system in the near future but haven't given any solid thought to what drives to use. at the rate drive prices are dropping for ssd, i may just get a few ssd's and chain them w/ lvm as a single drive, no raid. i haven't had any disk failures from anything except the cheap china 640G laptop drives i got last year and those happened within 2mo of purchase

Matthew J. Harmon - Dec 2, 2012

+David Ford I've heard good things about OCZ but had had lackluster experience with their support, refusal to replace with the same or similar product - SDHC cards. My Samsung SSD is running along great and I'll be looking into their on board AES-256 here today.

The laptop drive you mention sounds like the old IBM DeathStar drives. For two years I kept a tally on the right side of my whiteboard wall anytime I would hear about someone in my department hearing the "click of death", in two years we had over 200 replacements including several systems multiple times.

David Ford - Dec 2, 2012

nod my 1TB root drive on Scott is a WD10EARS. nothing particularly exceptional about it. not super fast, just average. got it because it's green

Matthew J. Harmon - Dec 2, 2012

My reason for going with ZFS / RAIDZ2 was several fold, self-repairing, compression and deduplication as chained operations and keeping greater than RAID10 reliability. Plus, the caching keeps frequently used data in RAM and on an SSD and long term / low use storage on the spinning disks. ZFSonLinux has become quite stable.

Matthew J. Harmon - Dec 2, 2012

Somewhat ironic that my RAID with a fault tolerance of two drives (before unrecoverable degradation) happened immediately. :-)

Matthew J. Harmon - Dec 2, 2012

3rd drive failing as I am on hold waiting for WD Red support regarding the other two drives. Maybe the 4th will fail before they answer! Starting dd if=/dev/zero on the 4th drive.

smartctl -a /dev/sdb | grep FAIL
SMART overall-health self-assessment test result: FAILED!
5 Reallocated
SectorCt 0x0033 133 133 140 Pre-fail Always FAILINGNOW 1967

I wish I came across this post earlier over at Hardforum:

"2 were DOA. 1 failed SMART after 10 minutes with reallocations. 1 failed DBAN after 3 passes. The drives came from 2 different resellers, and the 4 drives that gave up the ghost had almost consecutive serials dated 24 Jun 2012 from Malaysia."

My drives are 24 Oct 2012 from... Malaysia.

Matthew J. Harmon Dec 2, 2012

New MTBF for WD Red: 37 hours.

David Ford - Dec 2, 2012

yech :(

Matthew J. Harmon - Dec 2, 2012

I'm imagining the other Western Digital Red customers on hold along with me on this 24/7 "special" support line as skeletons with cobwebs holding a phone. Going on 2.5hrs of "We apologise for the delay in answering..."

Matthew J. Harmon - Dec 2, 2012

So after 3 hours on the phone waiting for Western Digital's special Red support line I decided it's not worth replacing these drives. All 4 have now failed and I really don't think I would have confidence in replacement drives from WD. Looking at Seagate Barracuda or Constellation maybe Hitachi.

Props to +Newegg for making refunding these as painless as possible. Quick chat support and a prepaid UPS shipping label for all 4 drives and a 3-5 day promise to refund.

David Ford - Dec 2, 2012

zero faith in wd red now

Matthew J. Harmon - Dec 2, 2012

I gave up after 5 hours waiting for someone to answer, 1:30pm till 6:30pm. I did however register the drives and report the defect, hopefully WD will do something about this horrible run of drives.

Matthew J. Harmon - Dec 5, 2012

3 days later, no response from Western Digital regarding my ticket.

Harmon, Matthew J. "Western Digital Dead Red Drives" Western Digital Dead Red Drives. Google+, 02 Dec. 2012. Web. 08 December 2014.

Matthew J. Harmon
Matthew J. Harmon -

Passionate security researcher, entrepreneur, consultant. Owner, founder, maker, mentor, teacher.

Loading Google+ Comments ...