Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Thanks for that link... So, basically consumer grade HDDs are generally quite reliable, but some are always going to fail, regardless of how they're used.

Just hope for a good one, and maintain reliable (and possibly numerous) backups.
 
Get a 1 TB Fusion and external SSDs via Thunderbolt for best performance / bang for the buck.

If you're not worried about having external devices sitting on your desk it's a no brainer: they're as fast or faster than the internal option, cheaper, and if they fail you don't need to ship back your mac.
 
I did a homebrew fusion drive on my MacBook and it failed...

Fusion is an interesting piece of tech which does work as advertised but it's too failure prone.

Er, Fusion didn't fail. You failed at implementing it correctly.

And can you point me to the data to show that Fusion is failure prone? If anything (anecdotal evidence) Fusion has been nothing short of a superstar. Unofficial usage may be a different story, but officially it seems robust.
 
Er, Fusion didn't fail. You failed at implementing it correctly.

And can you point me to the data to show that Fusion is failure prone? If anything (anecdotal evidence) Fusion has been nothing short of a superstar. Unofficial usage may be a different story, but officially it seems robust.

As mentioned, mine died within 2 weeks. I'm not interested in averages, just my experience, which in my case sucked.
 
SSD flash storage for iMacs are too expansive in australia (1,000 kanga dollars). When i come to purchase another iMac sadly i have to opt for a 3tb fusion drive for storage space and a cheaper alternative.
 
As mentioned, mine died within 2 weeks. I'm not interested in averages, just my experience, which in my case sucked.


My reply wasn't for you, OP.

It was for somebody trying to do fusion themselves and then complaining it didn't work properly and then made a broad brush statement that FUSION is unreliable.

Fusion is software that combines your SSD and your HDD. Your HDD failed. Fusion has nothing to do with it.

Some folks just want to Biitch about anything... Not you OP, the other bloke. You're entitled to be POed, your HDD broke after a few weeks.
 
The problem with the Fusion drive's reliability is that it's only 50% as reliable as a single SSD or HDD due to the coupling between these components.

Mathematically... that is incorrect. You are correct that reliability decreases.

/Jim

Except that having a Fusion drive reduces wear and tear on both components. So unless you have extensive statistics to back this up, it's just your assumption.

Actually... because Fusion has to continuously move data back and forth between the two devices... "wear" is actually increased. However, I think it is insignificant.

Don't forget that the fusion drive has double possibility to fail (two disks).

Mathematically, this is not correct. In order to compute the probability of failure... you must first take the product of the probably of each device NOT failing... which leaves you with the probability of the system NOT failing. Then you just subtract that from 1... leaving you with the probably of failure.

It is not simply twice the probably of failure because there are two components.

/Jim
 
Actually... because Fusion has to continuously move data back and forth between the two devices... "wear" is actually increased. However, I think it is insignificant.

It's not moving data back and forth continuously unless you're changing your usage habits on a daily basis. Data less used will always be accessed from the HDD after you've filled the SSD. Fusion isn't cache.
 
It's not moving data back and forth continuously unless you're changing your usage habits on a daily basis. Data less used will always be accessed from the HDD after you've filled the SSD. Fusion isn't cache.

I understand that. Still... compared to a volume that does not span two drives... there is increased data movement. In a non spanned volume... you write data to a HDD once... and it never leaves the device. With a fusion drive... you write the data to the volume... and it MAY move to the other device. It MAY happen more than once. Of course there may be unrelated data movement within any particular device. Some common examples are wear leveling on SSDs, or degrag on HDDs. These are all independent of fusion or not.

My comment was in regard to the statement made that fusion reduces wear and tear. I do not think that is true. I also think that it (at most) has negligible consequences.

The main point of my argument in this thread is that people continually describe failure rate as the product of the individual failure rates of each device (ex: since there are 2 devices... the failure rate is 2X). This is not mathematically true, and it overstates the real failure rate. Instead, to compute the failure rate properly... you must first take the product of the probability of each device NOT failing.

I am just pointing out that the fundamental methodology is being misstated. Nothing more -- nothing less. This is H.S. freshman math. It is not complex.

/Jim
 
The main point of my argument in this thread is that people continually describe failure rate as the product of the individual failure rates of each device (ex: since there are 2 devices... the failure rate is 2X). This is not mathematically true, and it overstates the real failure rate. Instead, to compute the failure rate properly... you must first take the product of the probability of each device NOT failing.

I am just pointing out that the fundamental methodology is being misstated. Nothing more -- nothing less. This is H.S. freshman math. It is not complex.

/Jim

while true, it does not make a huge difference unless failure rates or large or unless there is a significant positive correlation between the drive failures (e.g. if they are both more likely to fail due to some system issue such as a lot of heat).

But if the failure rates for both are 10%, and they are uncorrelated, then the overall survival rate is ( 0.9 x 0.9) = 0.81 . So the total failure rate is 19%, essentially twice the the single component failure rate.
 
while true, it does not make a huge difference unless failure rates or large or unless there is a significant positive correlation between the drive failures (e.g. if they are both more likely to fail due to some system issue such as a lot of heat).

But if the failure rates for both are 10%, and they are uncorrelated, then the overall survival rate is ( 0.9 x 0.9) = 0.81 . So the total failure rate is 19%, essentially twice the the single component failure rate.

I agree... and as you stated, it changes as failure rates of individual elements increase... or, if each element has a fundamentally different failure rate... as is the case with HDDs and SSDs.

The real reason I replied: I just get an allergic reaction when anyone incorrectly computes failure rates vs survival rates. It is in my DNA, and it makes me wonder how others see the world when I observe such a basic misunderstanding.

/Jim
 
Er, Fusion didn't fail. You failed at implementing it correctly.

And can you point me to the data to show that Fusion is failure prone? If anything (anecdotal evidence) Fusion has been nothing short of a superstar. Unofficial usage may be a different story, but officially it seems robust.

I did, really? You were there looking over my shoulder and saw how I set it up? You gleaned all that from one comment? Amazing, truly amazing. Actually no I did implement it correctly and it worked quite well for quite awhile until I used migration assistant to setup my Mac Pro, after that the fusion drive ate itself and rapidly degraded from there. Now please explain how using migration assistant would kill a fusion drive? You obviously know me so well tell me in detail what I did, what part of my implementation was at fault?

Don't make asinine statements and personally attack me all you do is look like an ass. My anecdotal eveidence much like your own tells me that a homebrew unofficial fusion implementation isn't as robust as one built by Apple.
 
I did, really? You were there looking over my shoulder and saw how I set it up? You gleaned all that from one comment? Amazing, truly amazing. Actually no I did implement it correctly and it worked quite well for quite awhile until I used migration assistant to setup my Mac Pro, after that the fusion drive ate itself and rapidly degraded from there. Now please explain how using migration assistant would kill a fusion drive? You obviously know me so well tell me in detail what I did, what part of my implementation was at fault?

Don't make asinine statements and personally attack me all you do is look like an ass. My anecdotal eveidence much like your own tells me that a homebrew unofficial fusion implementation isn't as robust as one built by Apple.

It was an unofficial install. That's why. You take your chances when you do things this way. Don't blame others, blame yourself. You know, take responsibility for your own actions.

You're the one who made the silly statement by writing that FUSION is unreliable based on your homebrew install of FUSION. That was a ludicrous thing to write.

BUt now we are in agreement - that a homebrew implementation isn't going to be as robust as one built by apple. As if that was never going to be the case.

Don't feel slighted. Nothing personal. You made a stupid statement and I called you on it.

Let's move on, shall we? I hope you can get your UNofficial Fusion to work fine in the future. Good luck.
 
3TB and 768GB is a big capacity difference. If you don't need all of that space, then the SSD option might be better... Or you can look into a smaller sized SSD and use the saved money on an external hd?
 
Sorry to start up this older thread but I can agree more with the fusion drive being poor, mine has failed 3 times now in 6 months, I never once had to take my old MacBook Pro to apple this has been back and forth to apple. It's getting to the point where I am just wanting rid if my iMac due to constant problems with the fusion drive
 
I couldn't disagree more regarding the lack of Fusion drive reliability. I have a 6 month old 2012 iMac that came with Fusion drive from the factory and its still running strong, not a hiccup. And I have a 2009 Mac mini with a DIY Fusion drive that is also running strong after three months, not a hiccup. From my experience I'm inclined to suspect user error (in the case of DIY) or hardware component failure for many that are having issues, niether of which is failure of Fusion itself.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.