Which 2 TB drive should I buy?

[rajsri]

Thanks... one more quick question, if I plan to use cache drive, does it's size need to match to parity's or other drive's size? I was planning to use 2TB for all the drives and 750GB for the cache drive. Is this fine?

[Superorb]

Thanks... one more quick question, if I plan to use cache drive, does it's size need to match to parity's or other drive's size? I was planning to use 2TB for all the drives and 750GB for the cache drive. Is this fine?

 

Cache drive can be ANY size, it is not tied to anything else.

[rajsri]

Thanks... I am going to build my FIRST unRAID server build over the weekend and shall post my results here (motherboard etc.,)...

[wsume99]

The rule of thumb related to cache drive size is that it should be large enough to handle the most data you would ever write to the array in a single day.  If you try to write more data to the array than can fit on the cache drive you will eventually get a disk full error.  Other than the first time you populate the array with data - which you should do before establishing parity BTW - I'd be surprised if you wrote more than 50-75GB of data to your array in a single day. But only you will know how much data you'll be writing. You can also partition your cache drive and use some of the space for installing apps like sabnzbd. You can basically use any old drive you have lying around for a cache drive.  If you don't have anything already I'd suggest something in the range of 160-320 GB.  500Gb would probably be WAY more than you'd ever need.  Just get whatever you can find the best deal on.

[furymaster]

So, I'm a *little* bit odd...

 

My cache drive is a 2TB drive - I use it until my server needs more space, then I'll add it to the array, that's why...

 

[wsume99]

^^That's not odd at all, that's just being wise.  I suppose that would be rule #2 for sizing a cache drive - make it as large as your largest drive so that it can serve as a warm spare.  That way if one of your array drives fail you can add it to the array a lot quicker than a brand new drive.

[aaronwt]

^^That's not odd at all, that's just being wise.  I suppose that would be rule #2 for sizing a cache drive - make it as large as your largest drive so that it can serve as a warm spare.  That way if one of your array drives fail you can add it to the array a lot quicker than a brand new drive.

 

This is the reason I made my cache drive a 2TB drive. I originally was going to use a 7200 rpm 500GB drive, but after reading up on teh cache drives, it made more sense to make it and the parity drive the same size for quick recovery in case there is a failure. So I'm using the Seagate Barracuda 5900rpm 2TB drives for the cache, parity and several of the array drives.

[Superorb]

So, I'm a *little* bit odd...

 

My cache drive is a 2TB drive - I use it until my server needs more space, then I'll add it to the array, that's why...

 

 

Several people do this, it's not a bad idea either.

[SSD]

If you have a smaller drive to use as cache, by all means use it.  If not, you are smart to buy a large capacity drive.  If your new drive is larger than one of your array disks, you can upsize the array disk and use the older drive for cache.

 

Somewhere earlier in this thread someone said that if your cache disk runs out of space, your copies to the array will fail.  That's not quite right.  There is a min free setting associated with the user share.  If that amount is free on the cache disk, it will attempt to copy the file there.  If not, it will pick another (array) disk to copy to.  If the min free setting is too small, and unRAID runs out of space copying the file, you will get the out of space error - even if there is lots of space on other disks.

 

The slit level can also dictate what disk unRAID needs to copy a file (i.e., to keep files in a directory together).  I am not sure how this works with the cache disk, but that could very possibly cause the cache drive to run out of space.

[hoek]

Just thought I'd mention something I've observed from 2 Hitachi 5k3000 drives.

 

After spinning up from either sleep or powered off the drive continuously clicks for a random amount of time.  Almost like it's randomly seeking, I can't quite describe it.

Can be a few minutes or much longer.  Then it goes silent.  After stopping, the clicking will very seldom start up again while powered on but almost always come back from sleep or after power cycling.

If while it's making this sound I read from it, it stops.  As soon as the read stops the sound comes back.

 

I actually returned the first one thinking that something must be wrong.  The second one does the same thing.

Both dated Dec 2010.

 

If you check this out

http://www.salvagedata.com/docs/failed-hard-drive-sounds/

And listen to the sound for  "IBM 40GB desktop hard drive with degraded media/heads rattles and squeals on spin up." it's very similar to that minus the squeals.

 

They did pass 1 preclear and parity sync with no issues.

I want to believe these are solid drives as I've had good experience with older 7200 Hitachi drives but this random clicking/rattling sound after spin up is not inspiring confidence so far.

I will likely return the second one.  Starting to think this whole "green" drive thing is not worth it.

[Superorb]

^^ Both my 5K3000 drives make a quiet tapping right after spinup, and their manufacture dates are several months apart including different firmware revisions.

[VampyreGTX]

It appears that my Hitachi 7200 just died (parity drive) last night. Still need to confirm it's the drive and not a MoBo or cable issue, but it does appear it's the drive (thread in the Gen. Troubleshooting section). There were no SMART errors leading up to it 24 hours prior. I haven't been able to access the drive to obtain SMART since, but will move it to my other PC to try and see how it functions there.

[ProfQ]

My Hitachi 5K3000 got lost in transit. Was insured, went ahead and bought locally another Seagate 2TB Green drive - ST2000DL003 Series.

[jaybee]

Hi there,

 

I thought it might be worth adding to this thread about a potentially serious issue with some Western Digital HDDs. A lot of people on these forums will already know about this and I have seen various threads discussing the issue. However, I could not see any clear mention of the issue at the beginning of this thread in the OP whereby possible issues with certain drives have been discussed. Although this can affect various Western Digital (WD) green drives, the main drive this seems to affect is the WDEARS drives judging from posts/googling.

I was thinking that this info should be attached to the OP to make people aware.

 

 

THE ISSUE

To summarize, out of the box these drives are set to park their heads after only seconds of drive inactivity. This causes a high LCC count to occur with some OS distributions where the drive gets woken up and put to sleep very frequently causing extra wear. Some people have commented that certain drives are only spec'd to have a lifetime of 300,000 LCC but in fact find their drives reaching this number within months.

Western Digital explain it better HERE and HERE and the text is pasted below incase the sites go offline. Included in this post are attached PDF and tool to download and use.

 

 

SOLUTION

Read the below and see attached files. There is a tool named WDIDLE3 you can use to flash the drive to eliminate this problem.

 

 

#######################################################################################################################

Description

 

This firmware modifies the behavior of the drive to wait longer before positioning the heads in their park position and turning off unnecessary electronics. This utility is designed to upgrade the firmware of the following hard drives: WD1000FYPS-01ZKB0, WD7500AYPS-01ZKB0, WD7501AYPS-01ZKB0.

 

CAUTION: Do not attempt to run this software on any hard drives other than what is listed above. Please make sure that the computer system is not turned off during the firmware upgrade. Doing so may damage the hard drive beyond repair and your data may be lost.

Supported Operating Systems

 

   * PC Compatible Computer (DOS)

 

Instructions

 

  1. Important: Please ensure that any important data is backed up from the hard drive and that the drive is removed from the RAID array (the array is offline) before updating the firmware.

  2. Download the wdidle3_1_05 file containing the firmware for the update.

  3. Extract wdidle3.exe onto a bootable medium (floppy, CD-RW, network drive, etc.).

  4. Boot the system with the hard drive to be updated to the medium where the update file was extracted to.

  5. Run the file by typing wdidle3.exe at the command prompt and press enter.

  6. The software will scan the Serial ATA bus and automatically update the hard drive that is connected and recognized. It may take up to two minutes to update the firmware of the hard drive.

 

     Critical: Please make sure that the computer system is not turned off during the firmware upgrade. Doing so may damage the hard drive beyond repair and your data may be lost.

  7. Once completed, a message will be displayed that the firmware has been updated. Once this message is displayed, remove your boot media and reboot your system.

 

Note: If you are having issues updating the firmware of a drive connected to a PCI Serial ATA controller or Serial ATA RAID controller, please install the drive into a system with Serial ATA connections directly on the motherboard (Onboard Serial ATA controller) and try again.

 

#########################################################################################################################

 

MUST READ: Western Digital is unable to provide support for the Unix/Linux operating systems outside of jumper configurations (for EIDE hard drives) and physical installation support.

 

Problem:

The Load/Unload counter for S.M.A.R.T Attribute 193 continues to increase under some distributions of the Linux Operating system and some Windows applications.

 

Affected Models:

WD20EADS, WD20EARS, WD15EADS, WD15EARS, WD10EADS, WD10EARS, WD8000AARS, WD7500AADS, WD7500AARS, WD6400AADS, WD6400AARS, WD5000AADS, WD5000AARS

Cause:

WD drives are designed to reduce power consumption, in part by positioning the heads in a park position (unloading the heads) and turning off unnecessary electronics, resulting in substantial power savings. WD defines this mode as Idle 3.

 

Some utilities, operating systems, and applications, such as some implementations of Linux, for example, are not optimized for low power storage devices and can cause our drives to wake up at a higher rate than normal. This can negatively impact some of the power savings associated with WD Green Power drives and artificially increases the number of load-unload cycles. The increase in load/unload cycles for a typical desktop user are within design margins (drive has been validated to 1 million load/unload cycles without issue).

 

Solution:

The number of systems using such applications and utilities is contingent on the types of system and applications being used. If you are concerned about your system we have provided three options depending on your system to disable the feature.

 

  1. Do not wake up the drives unnecessarily every 10 to 30 seconds or so, thereby gaining substantial power savings and eliminating excess activity. Increasing logging to every 2 minutes would result in 525,600 minutes per year or 262,800 cycles per year. Increase to 5 minutes and cycles would not even be a factor.

 

     a. Linux users: Decrease the logging message

         i. Examine your /etc/syslog.conf file for unnecessary logging activity and to optimize its performance. If you don't want to log any system activity, consider disabling syslogd and klogd entirely; or, at the very least, minimize the amount of logging your system performs. You can also prefix each entry with the minus sign (-) to omit syncing the file after each log entry. This will log anything with a priority of info or higher, but lower than warning, to /var/log/messages or /var/log/mail without needing to sync to disk after each write. Since we want to keep all messages with a priority of warning, this will be logged to a different file without disabling disk syncing (to prevent data loss in the event of a system crash).

 

            1. *.warning /var/log/syslog

 

            2. *.info;*.!warning;mail.none -/var/log/messages

 

            3. mail.info;mail.!warning -/var/log/mail

 

         ii. Another item to be aware of is the -- MARK -- messages that syslogd( writes. This will affect your hard drive inactivity settings. You can simply disable this by running syslogd( with:

            1. if [ -x /usr/sbin/syslogd -a -x /usr/sbin/klogd ]; then

            2. # '-m 0' disabled 'MARK' messages

            3. /usr/sbin/syslogd -m 0

            4. sleep 1

            5. # '-c 3' displays errors on console

            6. # '-x' turns off broken EIP translation

            7. /usr/sbin/klogd -c 3 -x

            8. fi

     b. Modify OS power management timers in control panel

 

  2. Disable Advanced power management using standard ATA command (Uses more power as turns off all low power modes but results in no load/unload cycles)

        1. Linux users add following (hdparm -B 255 /dev/sdX where X is your hard drive device). ATA users can disable APM usually controlled via BIOS and/or OS.

  3. Set Idle3 to max time (effectively turns off load/unload power saving feature thus will use more power) per below link.

 

Most of our customers do not disable the advanced power savings features. In addition to the options provided above we have also provided a utility. This utility will modify the behavior of the drive to wait longer before invoking Idle 3 mode. By disabling this feature your drive will consume additional power during periods of inactivity. This update is described in WD's Process Change Notice PCN 2579-701324-A02 (see attached PDF file).

The following article contains the download location: Answer ID 3263: The S.M.A.R.T Attribute 193 Load/Unload counter continue to increase for the WD RE2-GP SATA II hard drives=

 

MUST READ: Western Digital is unable to provide support for the Unix/Linux operating systems outside of jumper configurations (for EIDE hard drives) and physical installation support.

 

#########################################################################################################################

 

 

There are also some other main threads I have listed below that discuss the issue:

 

http://lime-technology.com/forum/index.php?topic=5992.0

http://lime-technology.com/forum/index.php?topic=9584.30

http://lime-technology.com/forum/index.php?topic=9851.0

http://www.xlr8yourmac.com/tips/Disable_WDGreen_HeadParking.htm

http://forum.qnap.com/viewtopic.php?f=182&t=28609&start=40

http://forum.synology.com/enu/viewtopic.php?f=124&t=20521&start=30

 

 

 

It seems that now we have issues with Samsung F4 firmware, Western Digital Jumper issues (to be aware of atleast, discussed in OP) and now this WD issue. It seems more than ever that the Hitatchi 5k3000 is a good drive to go with and works out of the box no fuss. Here in the UK you get 1 year less warranty with Hitatchi though and they are becoming discontinued as WD are buying Hitatchi out.

 

wdidle3_1_05.zip

[jaybee]

I could not attach the PDF in the same post as it was too large so here it is attached.

 

WD_Green_Power_EP250.pdf

[Rajahal]

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.

 

I personally don't believe that the WDIDLE3 tool should be used unless one actually sees an issue with their drive.  The tool messes with the drive's firmware and forces it to wait longer before parking the heads.  This reduces the drive's "green" benefits.  I prefer to allow my drives to be greener than to preemptively mess with their firmware.  Basically, I trust my drives as provided by the manufacturer until they prove themselves otherwise.

 

This is the same reason that I have not and will not buy one of the Samsung F4 drives.  Even though I understand they now work reliably with the JP1 firmware update, I don't trust a drive that requires a firmware update to work correctly.  While I haven't personally seen an F4 fail, I've read enough reports of issues with the drives on these and other forums that I've lost faith in them.  I've seen far, far fewer reports of problems with the WD EARS drives.

 

Another case of my stubbornness lead to my being bitten by the Seagate LP CC34 firmware.  I used a 2 TB Seagate LP drive with the stock CC34 firmware for several months before I saw it start failing SMART.  I RMA'd the drive and got a refurb drive as a replacement.  I immediately upgraded the firmware to CC35 (and was disappointed that it was still shipped with the faulty CC34 firmware) and have been using the drive as my parity drive ever since.  I later bought a second 2 TB Seagate LP that was on a sale I couldn't resist ($65 at time when most 2 TB drives were going for $90 - $100).  This one I immediately updated to CC35 and then ran it though preclear.  It failed preclear, so I RMA'd it.  I got another refurb drive and again updated to CC35.  This second refurb drive has also been performing reliably as one of my test drives.  Even though the two refurb drives I now have haven't been problematic, I somewhat regret purchasing them in the first place.  Both drives had to be RMA'd for one reason or another, and now I've ended up with two refurb drives.  While the refurb drives do work properly, if I ever wanted to resell them I wouldn't be able to get as high of a price as I would for non-refurb drives.  The great sale prices I found on each drive were also mitigated by the time and money I spend on RMAing both drives and messing with the firmware updates.

 

Well, that was a bit of a ramble, but the point of it all is that I believe that one has to look at a drive's use in the unRAID environment before applying a 'fix' that may or may not help in any way.  Again, if there's some large number of unRAID users that I've somehow missed who have been bitten by not applying the WDIDLE3 fix to their WD EARS drives, then I'll be happy to add a note about it to the OP.  However, until I've seen evidence to the contrary, I would prefer to not advise users to mess with their drive's firmware unless absolutely necessary.

[jaybee]

OK point taken and that explains it all nicely and also why it was left out.   I agree about flashing firmware. Personally I'm completely immune to worrying about firmware updates. The amount of firmware updates and dodgey hacking around of hardware I have done in my time, and I don't think I can ever remember bricking something via a bad flash. It's always been a physical thing I've done to break stuff! But yes thanks for the explanation and sorry to have caused perhaps unecessary alarm. I think for me, everything is pointing to getting the Hitachi drive given that is the top recommended drive in the OP and seems to just work out of the box. I think the lack of an extra year warranty isn't such an issue given that it has the most chance to go on working with minimal issues.

The only thing I notice, is that it has a spin speed of 5900 rpm and doesn't appear to be quite as green and power efficient perhaps as the WD EARS, however, I think they are all much of a muchness in this sector and differences are minimal, it normally just pays to get the cheapest drive at the time. For me, in the UK atleast, the 2TB drives are all within £10 of each other the WD EARS, Samsung F4, Seagate ST2000, Hitachi 5k3000. I notice that the Hitachi is NOT an advanced format drive, which I don't think matters at all, please correct me if I am wrong?

The Hitachi also has the added benefit of being the best performer of all the above drives, although I know that this is the least important factor really if using other drives that are going to be slower in the array...weakest link and all that.

 

 

 

[Rajahal]

I agree, the Hitachi is definitely the best choice of 2 TB drive at the moment.  I believe that the Seagate LP is the only 5900 rpm drive on the market, whereas the Hitachi, WD EARS, and Samsung F4 are all 5400 rpm.  Of course I can't confirm that as the Hitachi page is decidedly vague on the issue:

 

http://www.hitachigst.com/internal-drives/desktop/deskstar/deskstar-5k3000

Rotational Speed (RPM):  CoolSpin

Wow, thanks Hitachi, REALLY helpful

 

Anyway, I'm fairly confident that the drive spins at 5400 rpm but achieves better performance due to having higher areal density.  If you have an evidence to the contrary, let me know.  It is true that the WD EARS is the most power efficient of the bunch - it consumes about half as much power as the other drives.  In terms of cumulative differences we are talking about mere dollars (euros, pounds, whatever) per year of power savings, but if that is important to you then the WD EARS can't be beat.  I figure that in the long term they will all just about cancel out, since the Hitachi drives are a bit cheaper up front and the WD EARS will save you small amounts of money in the long run.  After a couple of years you'll want to replace them all anyway.

 

You are correct that being advanced format doesn't matter.  You can set-and-forget unRAID's option of '4k MRB aligned' for any drive and you'll get optimal performance out of every drive, regardless if it is advanced format or not (with the sole exception of a WD EARS with a jumper installed).

 

...I think they are all much of a muchness...

 

That some kind of British slang?  I have no clue what that means.... 

[jaybee]

 

That some kind of British slang?  I have no clue what that means.... 

 

It means to say, I think all of the 2TB drives are all very similar in terms of performance/power consumption. Yes a British saying. Like a slang term so as to say, they are all pretty equal and similar.

[Rajahal]

Thanks for the explanation....I want to say I've heard that before while traveling over there, but I can't be sure.  Traveling in the UK seems to inspire the imbibing of many ales, so I can't be sure of much that I've seen or heard over there

[Superorb]

Is it true that using the WDIDLE3 utility will void the warranty on the WD20EARS drives?

[Heretic]

you could always reset them no?

 

ive used wdiddle on all my ears drives, altho i  also believe this is more a problem for a linux  system drive that writes data to the disk at certain short intervals causing the heads to park ans startup way too many times. this should not affect an unraid server that is either reads writes or does nothing for long periods of time.

that being said if you use one of the unraid disks for an addon programs that write data to an log it might be worth using wdidle or use a differnt brand disk.

[dsj419]

Hey guys,

 

So I've been working on getting all the parts together to build the 4224 beast build 24 HD array as per spec'd out in the recommendation thread. I'm going to exactly duplicate it, but not quite sure about the HD's right now.

 

I was going to try and get 23 5k3000's and a 7200rpm cache drive, but i've just come accross a deal and wondering what you guys think and if its worth it:

 

I can get 24 Western Digital RE4-GP WD2002FYPS 2TB 64MB Cache drives for $4000.00 usd.  Should I pony up the money and get the faster drives do you think? is it going to make that much of a difference? This is only going to be a media server for 1080p bluray movies and 3D movies.

 

Let me know! Awesome forum!

 

 

*EDIT* I guess the other side to this is I only have about 8 TB's of movies right now. I do plan on expanding fast on that which is why I was going to just get the full beast rack, but I guess should I bother? It may take a year to fill up all those drives. Am I better getting all the components then just adding more hard drives as needed? All I have to do is plug them in right? as long as I start out with a 2 TB parity and a 2 TB cache and at least 1 storage drive I can plug them in at any time?

 

Thanks!

 

[Superorb]

^^ There's no benefit to having 7200rpm data drives. In fact, the added heat and power draw is an unwanted side effect of 7200rpm drives. Also, you don't need a 2TB cache drive unless you plan on using it as a warm spare, but with 22 data drives I'd recommend it.

 

This is how I'd do it: Get started with a 2TB parity and 5x2TB data drives to fit your current 8TB of files that need to be moved to the server. Then, when you have around 1TB free space remaining I'd order another drive since it'll take a while to get the drive as well as to preclear it. Make sure you preclear all drives, even the parity drive, before you put the server into service. I run 3 cycles of preclear on every new drive and 2 preclears on older, known good drives. You also might not want to buy a large batch of the same drives from the same vendor, this way you avoid getting drives from the same batch.

[dsj419]

^^ There's no benefit to having 7200rpm data drives. In fact, the added heat and power draw is an unwanted side effect of 7200rpm drives. Also, you don't need a 2TB cache drive unless you plan on using it as a warm spare, but with 22 data drives I'd recommend it.

 

This is how I'd do it: Get started with a 2TB parity and 5x2TB data drives to fit your current 8TB of files that need to be moved to the server. Then, when you have around 1TB free space remaining I'd order another drive since it'll take a while to get the drive as well as to preclear it. Make sure you preclear all drives, even the parity drive, before you put the server into service. I run 3 cycles of preclear on every new drive and 2 preclears on older, known good drives. You also might not want to buy a large batch of the same drives from the same vendor, this way you avoid getting drives from the same batch.

 

Thanks for the recommendation. So if I get all 5k3000 hitachi's, what about the cache drive? I see a lot of people talk about getting a 7200 rpm one.. not needed? i'd prefer to at the very least get the 2tb parity and a 2tb cache so I know i'm covered down the line.. I just want to make sure I get the right ones.

 

Thanks!