Western Digital Releases 5TB and 6TB Red and Green Drives

[Jomp]

http://www.anandtech.com/show/8273/western-digital-updates-red-nas-drive-lineup-with-6-tb-and-pro-versions

 

Newegg has the 5TB at 249.99 and the 6TB at 299.99.

[garycase]

Very nice -- it'll be interesting to see the performance figures for the 6TB version ... should be notably faster than the lower-capacity versions, since it uses platters with 20% higher areal density.    Won't help with transfers across the network (all Reds are already faster than network speeds) ... but a server with all 6TB Reds should have very high parity check speeds.

 

[Jomp]

 

Ohh, 5TB and 6TB greens too!

 

5Tb for 229.99 and 6TB for 269.99. Two year warranty vs three on the reds.

[Jomp]

Very nice -- it'll be interesting to see the performance figures for the 6TB version ... should be notably faster than the lower-capacity versions, since it uses platters with 20% higher areal density.    Won't help with transfers across the network (all Reds are already faster than network speeds) ... but a server with all 6TB Reds should have very high parity check speeds.

 

Ask and you shall receive 

 

"The WD Red 6TB is quite a surprising solution and if we analyze our test results, I can really see where this purpose built solution will fit in. First off, it is obvious this drive does not really care for workload testing, as it did not perform too well in any of that, nor did it perform too well with random read testing.

 

Where it did fare quite well is where it counts. The WD Red 6TB was the quickest in both sequential read and write by a good 30 MB/s, and at the same time had the lowest latency in both tests as well, coming in at 23ms over the 28ms turned in by the Seagate NAS 4TB and WD Red 4TB."

 

http://www.tweaktown.com/reviews/6535/western-digital-red-6tb-nas-consumer-hdd-review/index.html

 

[garycase]

Very interesting results -- but mathematically inconsistent.

 

The faster sustained transfer rates are as expected ... the 3 & 4TB units tend to provide ~150MB/s, and since the 6TB has a 20% higher areal density, you'd expect about 20% more data in the same amount of rotation ... 20% of 150MB/s = 30MB/s, or just what they found in their testing.

 

HOWEVER ... the lower latency doesn't make sense.  Latency is driven purely by the rotational speed of a disk ... the average latency time for any disk is 1/2 of the time it takes for a single rotation.  Size doesn't matter, nor does areal density.    Not sure how they measured the latency, but it doesn't make sense that the 6TB would have lower latency UNLESS their tests are skewed in some way that the results are perturbed by the 1.2TB platters.  I'm fairly sure the 6TB drive does not rotate faster than the 3 & 4 TB units.

 

 

[jumperalex]

I would assume that is in reference to the new pro drives that are 7200 rpm

 

Sent from my XT1060 using Tapatalk

 

 

[garycase]

I would assume that is in reference to the new pro drives that are 7200 rpm

 

Sent from my XT1060 using Tapatalk

 

No, the review is referring to the 6TB drives (the Pro series maxes out at 4TB).

 

[WeeboTech]

The link says... The 6tb has 5 platters. That could account for it.

[garycase]

The link says... The 6tb has 5 platters. That could account for it.

 

No, that simply means there's 1.2TB/platter.  The average latency is still 1/2 of the time for a single rotation of the platter.  The 1.2TB/platter clear explains why the sustained transfer rate is 20% higher than the 4TB units ... the platters have a 20% higher area density.

 

[c3]

Those latency numbers are not drive rotation latency, but IOP latency. Just shows the queue algorithm being used on the 64mb.

[garycase]

That makes sense.  Interesting that they've modified the algorithm ... the 4TB units have the same 64mb cache.    In fact, I was focusing on the difference and not the actual numbers or I'd have thought of this -- clearly the latency numbers are both far higher than rotational latency would be ... assuming 5400rpm the rotational latency would be 5.5ms (1/2 of the 11ms rotation time).

 

 

[mr_cb7]

will these  sizes work with the latest version of unraid?

[c3]

yes, 5tb and 6tb work with unRAID

[SSD]

will these  sizes work with the latest version of unraid?

 

There is no technical reason that drives up to 16TiB (max size of an RFS formatted drive) should not work in unRAID.

 

But until they are tried we won't know for sure. Controllers sometimes have firmware bugs. I had issues with a 5T Seagate (pulled from an external) that I think were a deliberate attempt to stop users from pulled drives from USB enclosures. Users have gotten 5TiB drives working, I have not heard of anyone running with 6TiB yet.

 

But again, in theory, they should work.

 

For anyone worried about the 16 TiB limit of RFS, this limitation will soon be lifted as unRAID will support XFS and BTRFS. These drives can be as large as 16 EiB (more than a million times larger than RFS). Imagine how long a parity check would take with an array of those puppies.

[merlyn]

will these  sizes work with the latest version of unraid?

 

There is no technical reason that drives up to 16TiB (max size of an RFS formatted drive) should not work in unRAID.

 

But until they are tried we won't know for sure. Controllers sometimes have firmware bugs. I had issues with a 5T Seagate (pulled from an external) that I think were a deliberate attempt to stop users from pulled drives from USB enclosures. Users have gotten 5TiB drives working, I have not heard of anyone running with 6TiB yet.

 

But again, in theory, they should work.

 

For anyone worried about the 16 TiB limit of RFS, this limitation will soon be lifted as unRAID will support XFS and BTRFS. These drives can be as large as 16 EiB (more than a million times larger than RFS). Imagine how long a parity check would take with an array of those puppies.

 

45 hours into preclearing two - 6 tb WD reds.  I am preclearing them both simultaneously.

currently 52 percent complete on step 10.

 

I will know if they both preclear ok by tomorrow... Then another day or so to set one up as parity.  Then will pull oldest, smallest capacity drive and rebuild that drive with old parity drive after preclearing that.  Will take like a week for the whole filter down process.

Will tell you how it goes from there.

 

I did notice that my bios recognized it as a 2 tb drive  but unraid recognised it as a 6tb.  Most Interesting my 4tb drives show up as 4tb drives in bios.  Silly BIOS.

 

my setup

Tamsolutions 4u rackmount supermicro servers that the forums here were raving about oh so many months ago.  I picked up the intel variant.

running unRAID Server Pro version: 5.0.5

 

merlyn

 

 

[garycase]

VERY strange !!    In general, if a controller works with > 2TB drives, it will work with ANY size drive up to the Exabyte size limit of the current structure.

 

If a 4TB drive is shown correctly, I'd expect a 6TB or larger drive to show correctly as well.    This much be a limitation of the BIOS code -- NOT of the controller on the system.  And the controller's all that really matters, since no modern OS uses the BIOS disk access routines.    In any event, if UnRAID is showing it correctly, it'll work just fine.

 

[merlyn]

VERY strange !!    In general, if a controller works with > 2TB drives, it will work with ANY size drive up to the Exabyte size limit of the current structure.

 

If a 4TB drive is shown correctly, I'd expect a 6TB or larger drive to show correctly as well.    This much be a limitation of the BIOS code -- NOT of the controller on the system.  And the controller's all that really matters, since no modern OS uses the BIOS disk access routines.    In any event, if UnRAID is showing it correctly, it'll work just fine.

 

I was in a hurry to get the pre-clear started and really didn't investigate further.  All i know is when i ran the preclear it recognized the correct 6tb size.  I won't know if unraid will show 6tb since i never tried to add it yet to the array.

 

Once the preclear is done i can reboot and check into it further.  I assume i have a old BIOS  (newest is 2011) who knows what version i am running.  cannot stop the array since it currently has family members using the array.  can you even try to add a drive as it is being precleared just to see what size it is in unraid .  I would say you probably could but i really dont know

 

motherboard is a supermicro based  X7DBE-X

i have 3 SAT2-MV8 Raid cards who knows what bios they are running cannot check while the preclear is still running. 

 

I am sure not going to abort a preclear 47.5 hours in to finishing just to check what raid bios version i am running.   

 

more info tomorrow when the preclear is done.  its currently at the last step and 3.5 tb into pre-reading (58 percent)

 

I am assuming all is well until it is proven not .

[SSD]

VERY strange !!    In general, if a controller works with > 2TB drives, it will work with ANY size drive up to the Exabyte size limit of the current structure.

 

If a 4TB drive is shown correctly, I'd expect a 6TB or larger drive to show correctly as well.    This much be a limitation of the BIOS code -- NOT of the controller on the system.  And the controller's all that really matters, since no modern OS uses the BIOS disk access routines.    In any event, if UnRAID is showing it correctly, it'll work just fine.

 

Not so surprising to me. Without drives of different sizes to test with I'd expect to find bugs.

 

But so long as the OS is seeing it correctly all is good.

 

I remember my first unRAID motherboard would report anything >=1T as 100.0G, but unRAID saw it fine.

[c3]

VERY strange !!    In general, if a controller works with > 2TB drives, it will work with ANY size drive up to the Exabyte size limit of the current structure.

 

If a 4TB drive is shown correctly, I'd expect a 6TB or larger drive to show correctly as well.    This much be a limitation of the BIOS code -- NOT of the controller on the system.  And the controller's all that really matters, since no modern OS uses the BIOS disk access routines.    In any event, if UnRAID is showing it correctly, it'll work just fine.

 

I think you mean PB, as in 128PB, the limit of LBA48.

 

As mentioned, there are lots of bugs found when larger drives come to market since the coders often make assumptions like "no one will even need more than 640k"... Indeed I found controllers (drivers/firmware) very happy reporting 4TB drives, only to actually work up to 4TB-1!

[merlyn]

both 6 tb drives pre-cleared fine

one took 63.43 hours the other 62.51 hours

 

Setup unraid to start rebuilding parity with the new 6tb drive.  currently reading at about 100mb/s  estimated time 1000 mins.

It will of course go much faster when it hits over 4tb since 4tb is the largest non parity drive currently in my 10 disk raid.

 

 

When i took a second look the controller bios is showing 1.9tb for the 2 tb drives 2.09tb  for the 4 tb and 6 tb.  So the 4tb have been  working fine for a long time so all is well.  Just as you all said its just a bug in the way it is displaying the size in the bios.  Works fine in unraid.

 

since i am running the server headless i never really noticed what it was displaying the 4tb drive as.  I assumed it was correct up to now.

 

in unraid under device status in the main webpage the size of the drive is coming up as 5860522532.

If someone gets a wd 6tb (non red) or a seagate 6tb  i would be interested to know the size it comes up as since this will effect which drive has to be parity (for future possible purchases).

 

merlyn

 

 

[garycase]

The size looks fine ... converting to "disk-drive-maker-ese" that works out to 6,001,175,072,768 => well over the advertised 6,000,000,000,000 bytes 

[redlaws]

Hi Guys,

 

I have finally seen some prices about the WD red 6tb about $399 in Australia.

 

I would love to consider building a box around say 4 6tb hdd.

so a parity check for a 2tb is around 8hr & a 4tb is about 16hr.

so 6tb I saw somebody quote about 60hrs is that right or just for test before using.

if that is the case then a ups would be a high pirioty for it.

 

Steve

 

 

[superloopy]

The 60'ish hours was for the preclear as i understand it ...

[garycase]

I'm sure the 60 hours was for a pre-clear.

 

A parity check on a system with 6TB Reds would likely take about 1.25 times what a system using 4TB Reds takes.    If the speeds were the same, you could reasonably expect 1.5 times; but the 6TB units have a 20% higher areal density, so parity check speeds will be faster.    In theory this could make them 20% quicker ... which would be about 1.2 times the 4TB time; but other overhead in the parity check would likely make it closer to 1.25 times.

 

[merlyn]

I'm sure the 60 hours was for a pre-clear.

 

A parity check on a system with 6TB Reds would likely take about 1.25 times what a system using 4TB Reds takes.

 

Yes that was me who quoted 60 plus hours for a pre-clear.

I did two simultaneously and the first disk took 63.43 hours the second took 62.51 hours.

 

I cannot help you on how fast just 6tb drives would be for parity since i am running multiple brand and size drives in my 10 disk unraid.  but it takes approx 1000 mins to do a parity check on my system now.

 

merlyn