I am slowly getting started with Unraid and like it so far. I like it a lot and am now ready to purchase a license and some drives 

 

I am looking for a bit of advice on which 2TB drives to choose. I have 3 WD AAKS drives which have all been at WD on warranty replacement. Also have a couple of 1TB greenpower which have not had any problems, but am somewhat upset with WD about the tripple failiures I got with the AAKS drives (and loss of 1.4TB worth of data). Plus, I see in the FAQ that the AAKS drives are crap and not worthy of use in UNRAID.

 

Seagate has always worked for me in the past, but I see most people prefer the WD green drives today. Also, the Seagate LP drives got some bad comments on Newegg about clicking noises. In my opinion Seagate handled the 1.5TB firmware problems very responsibly (I own one, but was not affected, a fact I was able to confirm on seagate's website).

 

I can still go either way I guess, and would appreciate to hear from anyone using the Seagate LP's or other 2TB drives. Specifically if someone has the clicking noises with the Seagate LP's, but also just any comments, really.

I started running unraid around June 2009.  I currently have 5 x WD2EDS and 3 x WD1001FALS.  Just did a parity check about a week or so ago, with no sync errors.  So far so good.  They're currently hooked to the Supermicro SATA controller.  I just recieved my second one and planning on installing that and plugging a 2TB HItachi from Fry's after I get it within the next few days...  Hopefully things will continue running smoothly...

I have 3xWD20EADS and 2xST32000542AS ... no issues to date.  I did have trouble with the Seagate 1.5TBs (ST31500341AS) in unRAID briefly freezing from time to time, but not with the 2TBs.

 

I've got a number of WD AND Seagate drives over the past few years. I'm still wrestling with the Seagate 1.5TB firmware. I didn't know about the latest SD1B and was on SD1A that worked well with the Drobos and for a couple of months on the unRAID before deciding to act up.

 

I've had failures from both WD and Seagate . That said, I've read Joe L.'s comments about only 2 types of drives (those that have failed and those that will) and couldn't agree more. Might as well just buy the drives you need when you need them and not worry about the brand. My 2c.

I have been looking hard at the current crop of 2GB drives and also find myself struggling to make a decission.

 

If money were no object, I would buy either the 2TB WD Black or RE4 drives.  

 

I recently did some performance comparisons with the new WD10EARS 1TB drive and didn't like what I saw.  You can check out my results at the end of this linked thread:

 

http://lime-technology.com/forum/index.php?topic=5076.135

 

I tried to quantify my 2TB options with the attached matrix.  You will see that some data points were unavailable from the manufacturers.  

 

Some things aren't quantified on this chart, such as the complaints on certain models - failure rate, noise levels, whining sounds (Hitachi, I think?).  

 

The Hitachis are starting to look like a good deal at half the price of the Seagate XT...

 

Will be putting my server in the basement so the noise is not that big a concern. Read and write speeds higher than the Green drives and Seagate LP's, access time as well.

 

Still thinking about it. The Deathstar scandal was so long time ago, under IBM ownership, and for me personally I only had 2 out of my 3 old 45GB drives fail in contrast to the WD AAKS drives, where 3 out of 3 failed (and one failed failed after coming back from warranty).

 

Oh, a question, I already have one WD20EADS as well. I need to check the capacity of the WD20EADS vs the Hitachis, as I would of course like to have one of the Hitachis as the parity drive if I go that way. WDC has put 2,000,398 MB in the data sheet, Hitachi just states 2 TB. Does anybody know the exact capacity of the Hitachi ?

 

I just added two WD20EADS (5400rpm) and transfer rates to those HDs max at 11.7MBs  while my Hitachi 2TB (7200rpm) max at 34-40MBs with parity on. Why is the big difference??

I experience faster transfer rates than your 11.7MB/s on my WD20EADS. Are you certain they're not EARS?

Both drives are WD20EADS. I should have mentioned that the 11.7 MBs is the max transfer rate when writing to the disk share. I get 45+ MBs when reading from the WD20EADS. I can write to my Hitachi 2TB (7200rpm) disk share at close to 40 MBs which is a big difference and won't require me to have a disk cache.

I experience faster transfer rates than your 11.7MB/s on my WD20EADS. Are you certain they're not EARS?

 

I only measured just over 15MB/s on a properly alligned EARS parity drive (with a very capable hardware platform).  I would say 12MB/s write performance is not hard to believe, depending on the controller and memory configuration.  The EADS has less cache memory than the EARS.

I just added a WD2EADS last night.  It's preclearing at about 80 MB/s.  Its the 1st drive on my 2nd SuperMicro SATA-8 Card...

 

Prior, when I perform a parity check, I think I was getting about 15-20 MB/s.

I will be moving to 2TB drives soon.

 

I've been doing research, and talking to my fellow storage guru (he's not using unRAID, though) on another forum. 

 

He's basically come to the conclusion that the Hitachi 7200rpm 2TB would be best for parity.  Yes, it runs warm, and is a little loud, but it's reliable, fast, and inexpensive.  Fine as a single drive in the system, but not recommended to have a whole case of them.

 

And for the other storage drives he recommends the Samsung 5400rpm 2TB.  It's cool-running, low-power, low-heat, and low-cost.

 

You can see that I'm partial to WD by my sig, but I've heard of high failure rates among the WD 2TB Green drives, and the WD 2TB Blacks are way too expensive.

 

Does anybody have any thoughts on all this?  And if I do go w/ Hitachi for Parity, and Samsung for storage will the Hitachi drive be as large or larger than the Samsung drives?  Any other issues particular to unRAID that I need to worry about based on these drives?

The Samsung drives report artificially low temps, by 5-10 C.

 

I'm having no issues with my 3 WD Green 2TB drives.

The Samsung drives report artificially low temps, by 5-10 C.

 

I'm having no issues with my 3 WD Green 2TB drives.

 

Interesting about the temps on the Samsungs.  Do you have a source for that?

 

What do you use as a parity drive?  One of your Greens?

 

Interesting about the temps on the Samsungs.  Do you have a source for that?

 

Mine reports a temp BELOW ambient... worthless.

The Samsung drives report artificially low temps, by 5-10 C.

 

I'm having no issues with my 3 WD Green 2TB drives.

 

Interesting about the temps on the Samsungs.  Do you have a source for that?

 

My Samsung is the same.  It consistently reports temps 5-10C lower than any other drive.  It is not in any special place in the array that gets more airflow, so I can only attribute it to the drive temp sensor.  It is very misleading!

 

Interesting about the temps on the Samsungs.  Do you have a source for that?

 

Mine reports a temp BELOW ambient... worthless.

 

It may have air conditioner inside it. You never know.  

 

So, are the temp sensors the only fault of the Samsung's? 

 

It seems you guys are leaning heavy towards the WD Green which is what I had planned initially anyway.

 

Thoughts on a 2TB parity drive?

I like my samsung drive, you just need to be aware of the temp sensor reading "problems."

 

Interesting about the temps on the Samsungs.  Do you have a source for that?

 

Mine reports a temp BELOW ambient... worthless.

 

It may have air conditioner inside it. You never know.  

 

Or, they have a peltier device inside.  (makes their drive run cooler, the competitor adjacent runs hotter)

 

Or, they simply have an "offset" of 10 degrees or so in the "raw" temp reading returned.

Or, they simply have an "offset" of 10 degrees or so in the "raw" temp reading returned.

 

Well I know that the temp sensors in Intel CPUs are more accurate near the max temps before throttling occurs (100*C) and less accurate the cooler you get.  They don't track linearly.  There really is no need for them to be accurate at lower temps as their purpose is to throttle or shutdown the CPU if it gets too hot.  They don't care if everything is just fine.

 

Perhaps the Samsung drive temp sensors are similar?

 

 

 

So, for storage the WD Greens and the Samsungs look the best, correct?  Just make sure to jumper the WD if you get the EARS, and ignore the temp readings on the Samsung.

 

And for parity is the general consensus still to use a 7200rpm drive?  If so, which is recommended?  And I from what I understand the parity drive has to be the same size or larger than the other drives down to the byte, so what are the exact byte counts on all the available drives?

In most standard cases (meaning no add-ons), a 7200 rpm parity drive will only help you if some of your data drives are 7200 rpm also.  When writing to the array, both the parity drive and the data drive have to be written to.  In this scenario, if either of the drives is slower (5400 or 5900 rpm), then the write will proceed at that speed, so the extra speed of the 7200 rpm drive will be unused.

 

That said, a faster parity drive can help in certain add-on environments.

In most standard cases (meaning no add-ons), a 7200 rpm parity drive will only help you if some of your data drives are 7200 rpm also.  When writing to the array, both the parity drive and the data drive have to be written to.  In this scenario, if either of the drives is slower (5400 or 5900 rpm), then the write will proceed at that speed, so the extra speed of the 7200 rpm drive will be unused.

 

That said, a faster parity drive can help in certain add-on environments.

 

OK, cool.  I may just stick w/ Green drives all around then, and only use faster drives for cache.

 

Edit:  What about during parity checks?  Does a faster parity drive help in that scenario?

Edit:  What about during parity checks?  Does a faster parity drive help in that scenario?

It will not help in a parity check unless the other disks are as fast.

The parity check will proceed at the rate of the SLOWEST drive involved.

I finally got around to replacing my parity with a 2TB Hitachi.  Temps are fine (see screenshot during parity rebuild). Speed dropped a bit though.

 

When I just had 4 Seagate 1.5TB 7200s, my parity check was ~92,000KB/s.  That dropped to ~87,000KB/s with the Hitachi as parity.  That being said, I have not noticed and performance degradation when writing to the array.

 

Internal speeds seem to be about the same:

 

/dev/sda:
Timing cached reads:   2200 MB in  2.00 seconds = 1100.43 MB/sec
Timing buffered disk reads:  360 MB in  3.01 seconds = 119.61 MB/sec

/dev/sdb:
Timing cached reads:   2194 MB in  2.00 seconds = 1097.39 MB/sec
Timing buffered disk reads:  368 MB in  3.01 seconds = 122.20 MB/sec

/dev/sdc:
Timing cached reads:   2134 MB in  2.00 seconds = 1067.90 MB/sec
Timing buffered disk reads:  364 MB in  3.00 seconds = 121.20 MB/sec

/dev/sdd:
Timing cached reads:   2178 MB in  2.00 seconds = 1089.54 MB/sec
Timing buffered disk reads:  366 MB in  3.02 seconds = 121.39 MB/sec

/dev/sde:
Timing cached reads:   2166 MB in  2.00 seconds = 1083.97 MB/sec
Timing buffered disk reads:  364 MB in  3.01 seconds = 121.04 MB/sec