Which 2 TB drive should I buy?

[Superorb]

^^ The cache drive only needs to be as large as the amount of data that you copy to the drive in any given day, so you can get away with a pretty small drive unless you're moving a ton of data at a time. Many people use a 500gb 7200rpm drive, but you could use a 2TB drive if you want a warm spare. I have a cache drive in my server, but I don't use it. It gets up over 40C easily, and I just don't need it. My transfer rates are around 35MB/s which is fine for me, especially since I'm moving 5gb or so at a time at most. Using all green drives will save you power and heat, and many of them are as fast or faster than 7200 drives except for random access which isn't really a problem in storage servers.

[perfessor101]

jaybee: I edited your post above to turn off smileys, which allows you to use "8" and ")" next to each other without getting this:   I did not change any of the post's content.

 

Anyway, I'm aware of this alleged issue with WD EARS drives, but I purposefully did not mention it in the OP because I personally don't believe it is a problem in the unRAID environment.  If you can show me a significant number of posts on these forums in which unRAID users have seen the LCC issue on their WD EARS drives, then I might change my tune.  When these drives are used in a regular desktop or RAID system in which they are spun up and down often then the LCC issue may be a significant concern.  However, as far as I've seen, this is not a problem in the unRAID environment.  I've been using several WD EARS drives (1 TB, 1.5 TB, and 2 TB) in my personal server for over a year now with no issues.  My drives likely spin up and down more often than most since I seed torrents from my server as well.  Even with that 'heavy' use, I haven't seen my LCCs skyrocketing.  If I did ever see my LCCs increase too quickly, then I would RMA the drive.  I prefer not to mess with a drive's firmware whenever possible, as the possibility of bricking the drive is all too real.

 

 

Hello Rajahal,

I was starting to build a new main computer, but thought I might first do some testing running the new unRaid 5beta on it.

I wanted a few drives for the unRaid 5 beta test server and shuffled the contents of drives around to clear up a few drives on 'storage', while I started testing things on 'tower'.  'storage' is in a cooler master stacker case with 11 bays and using cooler master 4-in-3 drive cages.  "tower" started out in an RPC-470 case and seeing how much better it ran unRaid I moved it into a RPC-4224 case after realizing the issue my drives were having with the LCC problem.  (also looking at replacing the drives one at a time as I could afford it was daunting ... unhook four drives ... remove a cage and replace one drive ... repeat six times about 30 minutes each)

I copied a few files to "Tower" and was trying parity build before switch to an early 5 beta and then I started trying to watch some recorded content from storage.  I could barely watch anything on it the recordings were stuttering like crazy. 

Went off and cooked up some dinner as I don't like flipping the switch on an unRaid server even if it is barely responsive. 

When I came back I switched off the power on 'storage' and started it back up again ... cancelled the parity check and waited for the drives to mount ... after they were all cleanly mounted I stopped the server, rebooted and started the parity check on 'storage'.  (some of my shares don't show up without rebooting)

 

I checked the information coming out of 'storage' under unmenu and checked the smart page from mymain and my load cycle counts skyrocketing had been causing the stuttering.  The buffer was just long enough to cause the drive to LCC every few seconds.

As far as I know I had no problem with the LCC's going nuts before moving drives around.

 

Most of the drives are still under warranty so I'm thinking after I have all the data securely off them, I'll run 100 preclears on each.

(Currently having issues with rsync and it's lack of support for unicode so it mangled my Itunes directory with it's all english content and other miscellaneous downloaded files going back to my attempts to convert video when 500 mHz single core processors were the norm ... blame it on Parkinsons Law ...)

 

Here are the LCC's on my WD 1.5 to 2TB drives (my 1TB drives didn't seem to be affected)

Parity Disk 472070, Disk2  240624, Disk3 237095, Disk4 270092, Disk5 363039 and Disk6 when it started failing the Data LifeGuard Diagnostic software was recently RMA'd.

 

Is it a problem for most ... probably not ... but until it got me ... I was not a believer

I'm usually the guy that gets lucky and gets things working most complain about and say don't work.  (Confirmation Bias?)

 

I still swear by my WD Drives still though they are 'easy' to RMA in Canada ... Advance Replacement I pay for return shipping.

With Seagate to get a cheap container to RMA with I have to get them shipped bulk from a far off land to reduce costs ... which costs in shipping and I don't need 15 RMA box sets (maybe eventually) ... and then I have to pay to ship it to Seagate and then wait for them to ship me a refurbished drive ... which will probably arrive COD ...

 

Bobby

[Superorb]

^^ You might want to rethink 100 preclears. That's over 100 DAYS of straight preclearing and is totally unnecessary. Three cycles is good, and 5 cycles should remove any doubt. Even 5 preclears on a 2tb drive is over 130 hours of clearing.

[perfessor101]

I would like to RMA those drives with over 250,000 LCC's, currently they are still in use ... and all other smart values are good.

 

I guess I should try running a long smart test on each drive again soon.

 

I can't remember where I read it or if the number is right ... but desktop drives are only guaranteed to be reliable up to 300,000 LCC's (or was that laptop drives?) so how long should I trust them for ... does that mean my 2TB drive is already out of warranty? even though it was purchased less than a year ago?

 

Sorry for the rambling ... time for sleep soon ...

 

Bobby

 

ps - for several months (actually since the first version of preclear_disk.sh and first new disk purchase after) ...  I have done three preclears on all new drives and on drives previously in use elsewhere I do one preclear unless I get 'funny' results.

 

The 1.5TB LCC drive I just RMA'd would have a random number of pending sectors (35 - 75) on each preclear with zero reallocated after each. WD DLG said read element failure.

 

although before Windows 7 came out I got a new Seagate 500GB boot drive and ran that through three preclears successfully before putting it into my windows XP professional computer.  It was dying within weeks and unusable before it was four months old and this was just after Seagate's Unspinnable firmware problem and Facebook had a few advertisers with corrupted .htaccess files so I thought it could be a malware issue.

 

[spinbot]

I'm looking at putting a 2TB Seagate (ST2000DL003) drive in for my parity.    It can either be plugged in to my main board (Gigabyte GA-MA74GM-S2) or my expansion Sata controller (AOC-SASLP-MV8).    My UnRaid is running 4.5.6 with 10 drives in it.  The Seagate drive appears to be an AES drive, yet no others in my array are.    Is this one of the "right" 2TB drives to use in my situation?  I believe I read somewhere that I should be updating to at least Unraid 4.7 as it will have a preclear that will work on the drive.  I recall before, you could move a jumper on the drives to disable the AES.  Is this still the case ?  Thanks

[Superorb]

^^ AES? Do you mean Advanced Format? Have you powered on the drive with the jumper already?

 

The Preclear script does not come natively with unRAID, so you can download it separately from the Preclear post in this forum.

[spinbot]

Advanced Format is what I meant...sorry about that.

 

The drive is in the mail and should arrive Tuesday, so i have not had my hands on it.

 

I can set it up which way makes the most sense going in as my parity drive.  I do have another 2TB drive, by Seagate, that i running as a data drive in my main computer (not my unraid server) that I will want to move over to the server shortly after.  The data drive I recall doing a firmware upgrade to.  I'm not sure if its advanced format or not (its sitting in a windows machine as I wanted to run it for a few months to see how it performed and make sure it didn't die ).

[Superorb]

^^ Basically, if you've already got the jumper on the drive it has to stay on there. But for a new drive, I'd leave the jumper off and use the newest version of Preclear. You can either use the "-A" (must be capitalized) option, or you can leave it as long as you set unRAID to format using "MBR 4K: Aligned" under the settings choices in the unRAID Main page. I'd recommend setting it to Aligned in the settings in either case. Starting a non-AF drive on sector 64 does not cause any harm.

[aoaaron]

thanks for making this thread! im new to unraid, servers and not that good at technology stuff.

 

I was about to grab 3x WDEARS drives but after reading this thread, I'll give them a miss so I don't have the headache (nor am i confident with the scripts etc.)

 

Are there any easy to configure drives available which you guys would reccomend for an UNRAID server acting as a storage to transfer date to XBMC via wifi and ethernet whilst also doing couch potato, sabnzb etc.

 

I'm from the UK so any drives under £80 would be great. im struggling to come to a decision for my first purchase.

[BRiT]

The EARS are easy to use. Do not jumper them and use Sector 64 MBR 4K-Aligned settings.

[aoaaron]

even what you said then sounds a little confusing . are there not drives available which are quiet and simple to configure?

 

i've browsed through many threads on the forum with WD EARS complaints.

[BRiT]

Frankly, if you think "Do not jumper them and use Sector 64 MBR 4K-Aligned settings" is too confusing, you should not be setting up anything yourself, much less a file server system. Pay someone to do the confusing tasks for you.

[madburg]

That is an option, there are great full time Builders on this forum that you can retain their services. If you dont know them already just ask.

[dgaschk]

The EARS are easy to use. Do not jumper them and use Sector 64 MBR 4K-Aligned settings.

 

This advise does not only apply to EARS drives. This is the same configuration for any drives in a new build or any new drive being added to an existing system. Use this configuration with any brand or type of drive that you select.

[Rajahal]

The EARS are easy to use. Do not jumper them and use Sector 64 MBR 4K-Aligned settings.

 

This advise does not only apply to EARS drives. This is the same configuration for any drives in a new build or any new drive being added to an existing system. Use this configuration with any brand or type of drive that you select.

 

Correct all drives (except for jumpered WDEARS drives) should use the MBR 4k-aligned setting.  The setting is on the unRAID 'settings' page near the bottom.  Just set it (to enabled) and forget it (as long as you don't use any jumpered WDEARS drives).

 

My personal favorite drives at the moment are the Hitachi CoolSpins.  The WDEARS are fine too.

[p1lot]

For building my unraidbox i want to use 4 old wd20ears i had been lying around. I want to get 2 to 3 new disks, should I get the same model or the Hitachi 5K3000?

 

 

Sent from my Nexus S using Tapatalk

[wsume99]

For building my unraidbox i want to use 4 old wd20ears i had been lying around. I want to get 2 to 3 new disks, should I get the same model or the Hitachi 5K3000?

You can expand your array with ANY drive you wish. You won't see any practical difference between using a WD20EARS drive or the 5K3000, or the equivalent Seagate or Samsung models.  That is one of the many benefits of using unRAID.

[SSD]

For building my unraidbox i want to use 4 old wd20ears i had been lying around. I want to get 2 to 3 new disks, should I get the same model or the Hitachi 5K3000?

You can expand your array with ANY drive you wish. You won't see any practical difference between using a WD20EARS drive or the 5K3000, or the equivalent Seagate or Samsung models.  That is one of the many benefits of using unRAID.

 

Go with the Hitachi.  Look here ...

 

http://lime-technology.com/forum/index.php?topic=10388.msg133845#msg133845

 

Unfortunately, unRAID does not have the ability to make all drives equally reliable.

[ProfQ]

I totally agree. Hitachis seem to be a simple choice ATM.

 

The Hitachis are plug and play in unRAID, and are proving themselves very reliable. 

 

[cbr600ds2]

Silly question but how often do HD's change/model numbers change in, say a year? 

[Rajahal]

Silly question but how often do HD's change/model numbers change in, say a year? 

 

2.5

 

Honestly, I have no idea.  The best I can do is cite Moore's law, which gives you the rough idea that any hard drive you buy today will be more or less obsolete within 2 years.

[cbr600ds2]

Hahaha...I was just curious because the drive names remain the same but its like the item number is off, slightly,. 

[Superorb]

Hahaha...I was just curious because the drive names remain the same but its like the item number is off, slightly,. 

 

Could be a small update, maybe design related, who knows. Companies rename things all the time for no good reason, even if nothing has changed and it's the exact same item.

[int13h]

Just a quick note : WD EARX drives seem to be OK. So far so good... Pretty much like a twin brother of previous EARS.

(4k-align, preclear with -A option, no jumper of course. )

[Compass]

I added a 2TB EARX last week, 1 round of preclear, 4K aligned, no jumper, added to the array, parity check with no errors, transfered just over 1TB music/movies, all good.