Difference between these two Hitachi 1tb Drives?

Finally looking to use a raid 0 config with an adata 64gb msata for intel RST in my Clevo P170EM. I actually do need 2tb total storage (sorry tiller) and wouldn't mind the speed boost from RST. I know there is a 2tb Samsung 2.5 HDD, but I'm wary of samsung HDD due to some failures in the past and the fact it is only 5400rpm. Hitachi drives have always done well for me. Also, while I do like the SSD + HDD setup I have in my old gateway, I'm not interested in going down that route for this machine.

I currently have this model as my primary drive
Newegg.com - HGST Travelstar 0S03563 1TB 7200 RPM 32MB Cache SATA 6Gb/s 2.5" Internal Notebook Hard Drive Retail Kit

However i just spotted this model
Newegg.com - HGST Travelstar 7K1000 HTS721010A9E630(0J22423) 1TB 7200 RPM 32MB Cache SATA 6.0Gb/s 2.5" Internal Notebook Hard Drive (Standard model)

Am I just dense and/or undercaffeinated? Is there a difference between these two sans price?

 

I think the 'kit' (Retail) may have an extra tidbit or two vs. the bare drive.

With 2 drive bays and an mSATA connector on your notebook; I would be looking at two 1TB EVO's for the drive bays (yeah: only 2.5" EVO models recommended because of no TRIM support on mSATA versions - see Anandtech review for more details) and for the mSATA slot? Save your money and leave it empty (or if you do populate it: use it for storage only).

No worries if you really need 2TB on your mobile system - but RAID0 with a 64GB mSATA is not even the last thing I would suggest for you.


(What is your usage model? Productivity or something else)?

 

I do enjoy the speed/performance of SSDs ( i have a samsung 830 256gb I use as a primary in my old gateway FX) but I am not a buyer at current prices. I'm willing to wait until we get to around $0.30/GB and hopefully larger than 1tb capacity.

For usage, I spend a week out of every month traveling. I'm typically without reliable wifi (either only unsecured public networks or good old hotel wifi which might as well be dialup).

I have close to 1tb of steam/gog/disc installed games. Despite having access to a 50mb connection at home, Steam downloads typically peak around 7-8mbs so installing/deleting/reinstalling isn't that efficient. I also rotate my games heavily - I typically play a different game every night depending on mood.

I have about 150gbs of music. Spotify/Pandora/XM radio apps/programs are all great, but again I'm dealing with low internet speeds and very spotty AT&T 3g service in the areas i travel to.

I have another 150gbs or so of comic books and book pdfs that I own. I don't own a tablet (don't care for them much)

Finally, i have well over 1tb of movies/tv shows and this is after pruning it down from close to 2tb. I'm a huge fan of pretty much all things David Attenborough and his series on BBC. Due to the quality, these are massive files ( 5 gbs or more in uncompressed format)

Finally, and least important, I have around 100gbs of work related material.

I've been making do with various external drives/docks. I've held off on doing a raid 0 + msata setup as it necessitates a clean install of win 7 (which i am admittedly overdue for) and quite frankly, I've been completely satisfied by the performance of the hitachi 1tb hdd. Boot up times are about double what the old gateway does with a SSD ( approx 30 seconds compared to around 60 seconds). Loading times/performance for games is also acceptable. I realize a SSD only or SSD + HDD setup would be much quicker, but for my own personal preferences I always like having a single C: drive to install everything to. I know it probably seems quite silly, but tis what I prefer.

My actual work workflow wouldn't benefit much from a SSD. It's data heavy (spreadsheets and database/CRM backups) and while I am sure I can improve latency/response, it's not slowing me down any.

I'm sure my gameplay could be improved with a SSD, but I'd be looking at either the evo 1tb or the crucial 960gb, and even then I would be stuck with externals. I could go the dual SSD route you suggested, but I'd rather spend that $700-$900 towards many, many other things.

I'm sure my Luddite HDD single drive setup wants are driving you mad, but given my work/game flows, I can't personally justify the cost for an all SSD setup nor do I want to save costs by purchasing sub par drives.

Given all of the above, do you still think a dual 1tb hitachi raid 0 + msata with Intel RST would be fruitless? I already have an adata 64gb msata and 1 tb hitachi, so the upgrade cost is simply another $70-$80 for an additional hitachi and some of my time. Any other upgrade solutions that would leave me with a single 2tb C: drive plus some sort of speed boost without going full monty 2.5 SSD? I've been looking around for months and this seems to be the best solution.

One final note is the P170em msata slot is only Sata II, not Sata III so a higher performance msata is still going to be bottlenecked.

 

First off, thank you for that detailed info. (Finally, a full and proper answer!).

Secondly, I agree that ~$1K on drives is less than what your current setup requires to meet your needs.

As for dual 1TB Hitachi's; yes. But not in RAID0 and definitely not with the mSATA SSD of only 64GB.

Even if you could get them seen as a single array; the capacity would be the size of the smallest drive x 3 (or about 178GB of usable capacity).


Even if you've already accounted for capacity hit with the smallest drive and mean just the Hitachi's in RAID0; still no reason to do it imo. The media files will not be read any faster, nor will saving them gain you much either vs. the ever present potential for one of the drives to glitch long enough for you to lose all your saved files on those drives.


If you're going to have an mSATA SSD installed along with 2 additional HDD's; keep them logically separate is my recommendation. RAID0 is a possibility without a reason in your case.

And since you can't make all three drives C:\ anyways... make a clean break and have a C:\, D:\ and E:\ drive from the start.


Come on in, the waters fine....

 

He can do RAID0 on Hitachi drives and then use the mSATA SSD as cache for the RAID via Intel SRT...

I have one Hitachi 7K1000 1TB in my P170EM cached by an mSATA Plextor 64GB. I use it as a dedicated drive for Steam. Works nicely, games' load times are reduced substantially.

 

Thanks to both of you. What Marecki said is what I had in mine. I wasn't too clear. My intention was to set up the twin hitachis as a raid 0 config with a fresh install of win 7. Once installed, enable intel RST and use the msata as a cache for it.

My only concern is the intel RST cache has a max of 64gb from what I have read, which is why i bought the adata 64gb. It appears to function by cataloging and recognizing the computer's most used programs (in my case steam and a few work applications) along with windows to speed up some operations. I really don't expect it to do much except reduce my boot up time into windows. Are those expectations in the ball park?

 

Intel RST caching of a boot drive will work exactly as an SSHD (i.e. Seagate Momentus XT). The main difference is that in the Momentus there is 4-8GB of flash, whereas Intel RST caching allows for using 64GB of flash.
OP your expectations are spot on, a 64GB cache SSD will accommodate most of your Windows and most of executables or libraries of the games you play and other software you use. Boot times and application opening times will improve dramatically.

 

I have played with Intel Smart Response Technology and my conclusions on my Windows based platforms mirrored what Anand came to with his Mac based systems; having an SSD cache drive smaller than ~250GB is money and time well wasted.

At that capacity, may as well go all SSD.

At lower capacities (like what Intel limits us to: 64GB), the performance consistency is hit and miss and imo not worth it.


I still don't recommend RAID0 at all - especially for a mobile system and especially not for the reason of having a single drive letter.

Also, an SSHD is not something to aspire for. And neither is RAID0 on a mobile setup.


I know I'll be ignored, but I have to state the facts. lol...


Keep in mind:

The problem of the drive letters will be transparent after the initial setup of the system: Steam games will go to one HDD and everything else will go to the other HDD and once installed, you'll never have to worry about them again.

I'm not sure why you're against a bigger mSATA though (unless it's just for price, which I can understand)?

With the O/S and your productivity Programs installed to an mSATA drive (even if it's SATAII), the system will be consistently much more responsive than the configuration you're considering. Not to mention more robust and finicky.

RAID: Dead, Invisible, Arrays, Randomly.


Possible?
I'm not sure if you can have a single SSD as an O/S and a caching drive (with partitions): but if you can, then even your Steam games can be accelerated more effectively than with all your data living in a single array.


Whatever you decide; Good luck!

 

It is possible to use a larger SSD. 64GB would be used for Intel RST, while the remaining capacity can be used as a normal partition, including an OS partition. 240-256GB seems optimal in such configuration. It's a pity, however, that the mSATA in P170EM is limited to SATA2 speed only.

Personally I would also refrain from using RAID0 in a laptop. I have to admit though, that I have my external storage configured as RAID0, just for increased performance. I am thinking of a 4-bay NAS/enclosure with RAID5 though, for increased performance AND security. Cost is some constraint though.

 

Marecki_clf,

Thanks for confirming that a larger SSD can be both an O/S drive and a caching drive simultaneously.

The performance that even a low end mSATA (which are even below 2.5" SSD's, performance-wise) drive offers above an HDD or even two great HDD's (like the Travelstars) in RAID0 is not to be ignored.

Especially when a good drive can be had for so cheap:

See (M500 mSATA 240GB $140 and less):
Crucial M500 Internal mSATA Solid State Drive, 240GB at Memory Express

 

Ahh, but the real problem is my steam library. Perhaps a better way to look at the situation is such

Project Budget: $80 (price of 1 new drive)
Current Materials: 1tb and 64gb
Project Scope: Take 1tb of critical data (the Steam Library) and provide the best performance. 1Tb figure is non-negotiable, as having to delete and redownload games isn't an option.

Ignoring the Intel RST and msata for a moment, I keep getting mixed feedback on RAID 0. Is this due to increased wear since both drives would be constantly in motion and therefor shorten lifespan?

I don't think temperatures would be a problem; I have yet to break 34C in 18 months. As far as potential loss of data, I am redundantly redundant. All my media/data/gamesaves are backed up WITH secondary backups, and I have all my really important stuff also backed up on good old DVD discs offsite. So if it just the risk that a drive will crash sooner with the potential for some increased performance, I'm in. After all, that's why you never forget to fill out the warranty card.

 

Warranty card on HDD/SSD's? What's that?

(Nobody has touched/accessed a HDD/SSD I've personally used even for a day).


The issue with RAID0 is that each additional drive you add to an array increases the chance of the array as a whole dying by more than 2x (for the first two drives).

Also, offsite storage won't help when you want your system available - sure, the data may be safe but sometimes that is not good enough.


What my biggest concern is: you're taking a stable setup and want to create a potentially unstable setup. Sure, higher performance (which I crave) but at additional risk (which makes no sense based on your needs).


I think we have presented all the possible options for you to consider.

 

This is incorrect. For each additional drive after the first, the probability of any of the drives failing increases by less than the probability of failure of a single drive.

The probability of failure is 1-(1-P)^n, where P is the probability of a single drive failing, and n is the number of drives in RAID 0.

For example, if the probability of a single drive failing within one year is 2%, the probability of at least one drive failing in a RAID 0 array of 10 drives would be just under 18.3%.

 

Qing Dao,

While your math may be correct, your assumptions are not. The HDD is not the only thing in an array; we've introduced other variables into the equation as well.

With your 10 drive array example; what you're stating is that if we had unlimited monkeys to play with unlimited RAID0 10 drive arrays, 1 in 5 (or 1 in 6, depending how optimistic you are) would be failing continuously.


I can't see that as a good argument for RAID0.



Here's a pretty good link to a RAID0 calculator that takes into account number of disks, age of disks and age of the array - with a pessimistic or optimistic outlook selector too.

See:
RAID 0 Lifetime


If we use the calculator to show the difference in the first year between a single year old disk and two, year old disks (Optimism: Optimist), we see the chance for the survival of the RAID as a whole goes from 92% to 85% in that first year. Even more eye opening is at the 4 year mark, the single year old disk has 75% chance of survival while two year old disks have only 56% over that same 4 year period.


Don't forget that some of our infinite monkeys on infinite arrays will also see total failure in the first bootup while some will also see impeccable operation for the next dozen years.


What I'm suggesting to Ajfountains is to not tempt fate. When the odds are stacked against you like the above examples show, minimizing your risk is more prudent than showing how fearless you can be.

Fearlessness is a great trait to have if you're defending your village against a lion or a pack of wolves; the possible rewards justify the risk involved (and doing nothing will still make you a meal for a carnivore).

But using that kind of bravado against cold hard facts is a little silly from my perspective.



But of course, I can only say this after I had a few good (and bad) years of playing with RAID0 myself.

 

If my math is correct, which you agreed it is, then it means that what you said before about the risk of loss more than doubling for each additional disk was wrong. That is all I meant by my post. I used a 2% failure rate and a RAID 0 array of 10 disks just to show how the equation works, not to tell people that it is great to run 10 disks in RAID 0.

Just so you know, according to that calculator, the "optimistic" (there is almost no difference between optimistic and pessimistic if you actually take a look at it) 1 year failure rate of a brand new drive is 2%, EXACTLY what I randomly chose to plug into my equation. Yet you somehow try to use that calculator as proof that my example to show you how the equation works is somehow wrong. Obviously the failure rate plugged into the equation has to be with regards to some unit of time. I just thought that was self explanatory. Over a great enough time frame, the failure rate is 100%.

Tiller, I'm not sure why you insist on plugging in 1 year old drives into the equation on that page. You would think that someone interested in buying new drives to put in RAID 0 would not try to go out of their way and purchase used drives.

Anyway, you have stumbled upon a lot more than you realized, the Google hard drive survey data. Google put it together to give a good gauge of how their drives (without stating what manufacturers and what models) had lasted in their servers. That site uses that information to determine the failure rates in the equations. These numbers really are a testament to the durability of hard drives. If you look at the "pessimistic" numbers, there is a 70% chance that any hard drive that Google uses in their servers will still be functioning after FIVE YEARS running 24/7. Also, after one year of 24/7 use, the failure rate of their hard drives is only 2%. That is astounding. There is no way that any drives in a personal computer will be used and stressed that much.

I know you tried to use that online calculator to try to make it look like RAID 0 is really dangerous, but those kinds of numbers really do nothing other than to boost confidence in a RAID 0 setup.

Check your statistics. Fail every year, not continuously.

 

It's okay, I understand that you don't understand statistics - but note that for the one drive at least that the OP will be using is about 1 year old (certainly not new, anyway), unlimited monkeys with unlimited arrays would be failing continuously (not yearly... sigh) and you still don't get that the statistics of an array failing as a whole is more than the statistics of simple multiple drives failing on their own.

I am trying to show by that calculator that the risk is not imaginary.

While 2% failure rate for a single new drive is quite good - that is not the expected failure rate over the lifecycle of the notebook (which is the point I'm trying to get across).

If the OP wants to be rebuilding his system from scratch (willingly or as a learning process) then sure, RAID0 is the way to go.

If however, the OP wants the fastest, most stable, highest capacity setup he can get with the least $$$ and the most sustained performance; those options have been discussed above.

 

You said that 1/5 of the arrays would be failing continuously. 1/5 would fail in the first year. That is different than saying that if you had an inordinately large number of arrays, there would be continuous failures. Anyway, I don't know what you are trying to get at by saying this. It has nothing to do with anything. If I have a trillion SSD's in RAID 1, there would be continuous failures as well.

If we assume that the rate of hard drive failure is independent of each other, then the failure rate of the array is actually statistically predictable using the failure rate of a single drive.

But if you have an actual reason as to why you think the failure rate of RAID 0 would be more than double the failure rate of a single drive, instead of just saying it is so and insisting that I don't understand some other underlying reason that you have yet to explicitly mention, then I am all ears.

 

Qing, I can't help if you don't read. Sorry.

My statements are accurate and if you can't get why a RAID0 array has a greater chance of failure than the combined statistical probability of two HDD's then someone else may be able to explain it better.

Thank you for trying to understand - but this forum on this corner of the web doesn't hold every single answer.

And just because I know something doesn't mean I have the time to explain it fully either.

 

So your final reply continues to be "just because tiller says so?"

Gotcha.

 

Why do people keep harping on about RAID0 more likely to fail. Any drive can fail, backup is the key.

The more drives a computer has the more chance one of one of those drives will fail. That is a if a computer has 2 drives in RAID0 or 2 separate drives the chances of a drive failing are the same so logically why not also harp on the users with 2 separate drives that they should only use one drive? And if you think only having one drive is going to make you immune from a drive failure, think again

 

If you have 2 separate drives and one of them fails, you only loose data on that drive; the data on the other drive stays intact. If you have 2 drives in RAID0 and one of them fails, you loose ALL your data. That is, of course, without putting any backup into perspective.

 

So are your saying backup isn't necessary if using one drive because the chance of failure is lower? or if you use 2 separate drives only one drive needs to be backed up or both don't need to be backed up because you only lose half your data? Perspective.comes into it a lot, I personally would backup all drives whether they are part of an array or not. Certainly wouldn't trust on a drive not failing unless it was data I was not interested in keeping. Too many drive failures over the years to think differently.

 

I am not saying anything like that Stop trolling please

Backups are needed and in professional environments they are a must.
Funny thing is that I don't have RAID0 in my laptop; I have my backups on a RAID0 array on external storage though ;-)

 

My intention is not to troll, sorry if it came over that way, just that with backups drive failure should not be a factor whether RAID0 or not and backups should be done if you value your data regardless if it is on single drives or part of RAID0. For myself the laptop has RAID0 and a single drive. Backups are done on single external drives, more than one drive with the backups.

 

The issue with RAID0 setups is not with regards to backups: it is to regards with how available you need a system to be.

If you have the option of using a computer or not (eg. gaming) and a glitch takes out your RAID0 configuration for a day or two, no problem.

If you are depending on this (possibly your only system) to be available always (work, school, etc.). then a glitch turns into a full blown emergency with dire consequences.

Whether you have backups or not. Time is money and RAIDO does not protect you from the non-availability of a system - again; whether you have a backup or not.


The few cases where high sequential speeds make a difference (video/audio) is also where more than likely more than one workstation is available and more than likely more than one of those workstations have access to the original source data that is being processed on one of the RAID0 setups.

We're not talking about simply losing data here. We're talking about losing a system, period.

If you only have a single system to use; RAID0 is not the answer when system availability is the question.

Also, if the glitch is software - then sure, a rebuild from a backup/image will get you up and running.

If the glitch is hardware - then only after messing with a backup/image rebuild will you know your system is still (effectively) dead.


The difference between these situations with a single HDD and a RAID0 array is that a HDD will usually give you some period of warning before it goes kaput.

RAID on the other hand depends on the platter fairies and where they put their magic wands that moment.


The interesting thing to me is that when I was doing RAID0 (multiple VRaptors was my favorite setup ), while the arrays would fail from time to time - the individual drives continued to be rock solid for years after I stopped using RAID0. This is the point. The array as a whole behaves differently than individual drives do with regards to dependability/stability/availability.


Backups to multiple sources (and created multiple times during the creation of the project - depending on the importance ($$$$) of the project) is a given (or should be).


Who knows? Maybe today's O/S, drivers, programs and/or HDD/SSD firmware play nicer with RAID arrays...

But all I know is that when I have a single system with me at a client's (which is rare; minimum 3 identically setup platforms are my standard 'availability' backups - if they all fail; I can't be blamed for trying to be ready for the job as promised), even when it is just running my accounting software and we're hammering out the last few billing details; I am more than a little on edge until the last byte gets uploaded to my personal cloud and/or to the client's servers and/or to my external drives (both SSD and HDD based).


A recent new client found out about the 'availability' side of computers the hard way: they gave me backups that were 10 minutes old of their billing system. Great! They also had to wait a week for a new system to be approved, ordered, setup and then get in contact with their software providers to allow their software to run on the new hardware.

The 'simple' problem of getting a new system (as they initially explained to me over the phone) turned into a mini-network with 3 identical systems with a dedicated NAS for their estimating/billing/procurement department (all using essentially the same base data).

And they weren't even running RAID0.


I don't know how much more clear this can be made:

1) backups are not to be questioned (ever; even if doing the 12th one that day).
2) RAID 1/5/6/10/etc. is not a backup, let alone RAID0.
3) a single system should be built with availability in mind over minimal performance benefits.
4) RAID0 has more points of failure than simply adding the probability of all the single drives failing individually.
5) multiple workstations with access to identical data allows for experimenting and using RAID0 setups - anything else is irresponsible imo.


The above is what I've gleaned from a few decades of playing with arrays and different RAID setups.


Anyone and everyone is welcome to recreate the RAID0 wheel though.

As long as it is understood what they're trying to accomplish.

 

I respect your findings Till and your higher than the average users needs. What I don't like is when people post RAID0 as if it's in imminent failure. Plenty of laptop manufacturers suppling RAID0 systems and for most people failure if it happens is going to be a PITA but usually not loss in revenue.

For critical time is money scenarios I would go even further and have a spare laptop / PC should it fail. Backups on their own would not be enough, more reasons than just the disk for failure.