Register for the iXsystems Community to get an ad-free experience and exclusive discounts in our eBay Store.

Desktop Drives

tomahawkeer

Junior Member
Joined
Oct 16, 2019
Messages
12
So I have seen in a few people's build signatures, they are listing Desktop drives as in use in some of their array's. Isn't this bad juju with ZFS / Freenas?
 

Newfoundland.Republic

Senior Member
Joined
Jul 2, 2019
Messages
368
I'd say it is up to your risk tolerance.

I have my FreeNAS using NAS drives (WD Reds) in a RAIDZ2 because I want the protection. That is why I back up using rsync to my old Synology NAS (mirrored drives) and then to another, external USB drive. NAS (or Enterprise like WD Gold, Seagate Exos, etc.) are made to be powered on all the time. (I'd stay with NAS drives and the performance of the Gold/Exos drives likely isn't needed in most cases.)
 

boyese

Neophyte
Joined
Dec 2, 2019
Messages
6
I'd say it is up to your risk tolerance.

I have my FreeNAS using NAS drives (WD Reds) in a RAIDZ2 because I want the protection. That is why I back up using rsync to my old Synology NAS (mirrored drives) and then to another, external USB drive. NAS (or Enterprise like WD Gold, Seagate Exos, etc.) are made to be powered on all the time. (I'd stay with NAS drives and the performance of the Gold/Exos drives likely isn't needed in most cases.)
Cheers - I cant afford the type of disk resiliance that you have but I'll look at getting another 8TB.
 

tomahawkeer

Junior Member
Joined
Oct 16, 2019
Messages
12
It's been several years, but I do know that I used WD green drives in an old Drobo 8 bay NAS, and they did NOT work well.

I replaced all of the greens with wd black drives, at the time I think they were considered enterprise desktop drives, and they worked great.

Those were eventually replaced with WD reds, which also worked great. The black drives that were in that array, are ALL still functioning in multiple workstations and are approaching 6 years old. They were used in the NAS for about 2.5 years non-stop

I realize that a Drobo isn't a freenas box, but they do have general functionality in common, so I would like to hear from those who are using desktop based drives, and get some feedback from them.
 

rvassar

Senior Member
Joined
May 2, 2018
Messages
516
I'm not sure, as I never buy them, but... I believe the difference is the firmware loaded on the drive controller. Specifically:

1. Multiple drive vibration reaction. Ala: The drive next to me just did a long seek, vs. I've been dropped, park heads now.
2. Degree of SMART diagnostic compliance.
 

LVLouisCyphre

Junior Member
Joined
Dec 22, 2019
Messages
16
So I have seen in a few people's build signatures, they are listing Desktop drives as in use in some of their array's. Isn't this bad juju with ZFS / Freenas?
Not necessarily as you expect the drives to fail especially with a RAIDZ2 or 3 which is why we use RAIDZ2 or 3 or have hot or warm spares. You have some fault tolerance in that configuration. You may also have to in a pinch build a NAS out of desktop hard drives or use a desktop drive for disaster recovery as a temporary drive. However, enterprise or NAS drives are designed for a NAS and should be used for that purpose but IMO it's not a NAS breaker or disaster.

The bad juju IMO is with the following;
  • Not enough ECC memory; this is probably the number one issue I've heard with ZFS, going cheap on memory or thinking they can get around it by having a large swap space.
  • Not burning in your hardware before putting it into production; Uncle Fester's guide makes sense to me. This includes memory and CPU upgrades and hard drive replacements. When you replace a hard drive, do the hard drive verification process in the guide before resilvering. You want the hardware to break before it's in production. A cascade drive failure is not a good thing during resilvering.
  • No UPS; even a dumb UPS with serial or USB monitoring to your NAS is better than nothing.
 

Alecmascot

Senior Member
Joined
Mar 18, 2014
Messages
375
I use mainly WD desktop drives in my NAS. They work fine after running wdidle.exe.
I have ages running from 25K to 75K hours.
 

HolyK

Ninja Turtle
Moderator
Joined
May 26, 2011
Messages
461
I have six WD-Greens in Z2 and they have ~2400days PowerOnHours (6.5 years, 24/7 operation). One of them recently got 7 Offline_Uncorrectable sectors pending for reallocation and ~500 of Multi_Zone_Error_Rate (signs of mechanical issues with either disk surface or heads). The other five are in good condition with zero errors/realocations, ...

Model Family: Western Digital Green
Device Model: WDC WD30EZRX-00DC0B0

ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
1 Raw_Read_Error_Rate 0x002f 200 200 051 Pre-fail Always - 0
9 Power_On_Hours 0x0032 022 022 000 Old_age Always - 57326
196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0
197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 0
198 Offline_Uncorrectable 0x0030 200 200 000 Old_age Offline - 0

I recently bought 6 new 12TB disks because i was about to hit the 80% fill level otherwise i would jut replace the one green and keep using the rest 5.

Friend of mine had WD RED completely blown up during burn-in. It just got screwed right away. Hunderds of errors, uncorrectable erros, pending sectors. He RMA'd it and got replacement. The replaced one had first reallocation event like 2 months after 3y warranty period ended. It was spreading rapidly so it got replaced. On the other hand i have two REDs with nearly 2k days and zero issues.

Soooo i would say there is no issue with either of these. It is more about the plain "luck" if you got the one which will run for 10 years without hiccup or the one which fails few days after warranty :/

What IS important (at least for me) is to ensure that the disks are from different batches or at least the number of disks from the same batch is less (or equal in wort case) than number or parity drives. Reason is simple ... if the disk failure is not caused by the plain "random" effect but due to some hidden manufacturing issue there is a risk that the disks from same batch will fail as well sooner or later. Having multiple disks failing at the (nearly) same time is nightmare ...
 
Top