Value System Build

[SushiEater]

theswede wrote:

I was only kidding to see if you catch it.

Right.

All I know on my 120GB drive in 2 weeks of doing panos I lost one year. I don't know what kind of writing was happening. All I know is the final result.

MTBF predicts this. You've been lucky with HDD's and had bad luck with an SSD. Statistics support that happening. It says nothing at all about either HDD longevity or SSD longevity; it takes a lot more datapoints than that to draw conclusions.

So you're right. That is all you know. And it tells you nothing about how a properly managed SSD will last compared to a properly managed HDD.

Jesper

So how exactly do YOU properly manage SSD?

All I did is to assemble many large panos together after copying all the files to SSD, copied them to hard drives after assembly and erased them from SSD.

Should I have done something different?

 

[theswede #100093]

So how exactly do YOU properly manage SSD?

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

All I did is to assemble many large panos together after copying all the files to SSD, copied them to hard drives after assembly and erased them from SSD.

There was no OS involved? No parameters set anywhere at all? This just happened, out of the blue? You're not providing much information, leading to me believe you're not really asking. You're just trying to make some kind of point from your anecdote.

Should I have done something different?

I have no idea, since I have no idea what you did to the drive.

Jesper

 

[SushiEater]

theswede wrote:

So how exactly do YOU properly manage SSD?

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

What is the point of having SSD if you use memory for the RAM disk? Other than that I have the same thing you have plus I turned off indexing.

All I did is to assemble many large panos together after copying all the files to SSD, copied them to hard drives after assembly and erased them from SSD.

There was no OS involved? No parameters set anywhere at all? This just happened, out of the blue? You're not providing much information, leading to me believe you're not really asking. You're just trying to make some kind of point from your anecdote.

For the life of me I don't understand what you are asking me to provide beyond what I already have.

I have OS and all programs and the scratch disk on the same 120GB SSD. About 60GB is free. I have decided to see if assembling panos will be any faster if all files are on the SSD. After 2 weeks of doing it SSDlife reported one year less of predicted life.

I have stated this before if you were to bother to read.

Should I have done something different?

I have no idea, since I have no idea what you did to the drive.

Jesper

 

[theswede #100093]

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

What is the point of having SSD if you use memory for the RAM disk?

Faster loading of files? Why is this a question?

Other than that I have the same thing you have plus I turned off indexing.

Well, there you go then!

I have OS and all programs and the scratch disk on the same 120GB SSD. About 60GB is free. I have decided to see if assembling panos will be any faster if all files are on the SSD. After 2 weeks of doing it SSDlife reported one year less of predicted life.

Ah, so no failure then. Then I must admit I don't see what the problem is? If HDD's could be as reliably diagnosed I'm pretty sure your beloved RAID 0 setups would get very scary readouts too. But they're just readouts, and have little to do with reality.

I have stated this before if you were to bother to read.

I was certain I'd missed something as what you said before was just a lot of ado over nothing. And it turns out it still was.

Jesper

 

[SushiEater]

theswede wrote:

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

What is the point of having SSD if you use memory for the RAM disk?

Faster loading of files? Why is this a question?

Because most people use SSD to load files faster or have enough memory so they don't need RAM disk.

Other than that I have the same thing you have plus I turned off indexing.

Well, there you go then!

I have OS and all programs and the scratch disk on the same 120GB SSD. About 60GB is free. I have decided to see if assembling panos will be any faster if all files are on the SSD. After 2 weeks of doing it SSDlife reported one year less of predicted life.

Ah, so no failure then. Then I must admit I don't see what the problem is? If HDD's could be as reliably diagnosed I'm pretty sure your beloved RAID 0 setups would get very scary readouts too. But they're just readouts, and have little to do with reality.

Did I ever say my SSD failed? But if I continued going the same way it would.

I have stated this before if you were to bother to read.

I was certain I'd missed something as what you said before was just a lot of ado over nothing. And it turns out it still was.

Jesper

 

[theswede #100093]

SushiEater wrote:

theswede wrote:

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

What is the point of having SSD if you use memory for the RAM disk?

Faster loading of files? Why is this a question?

Because most people use SSD to load files faster or have enough memory so they don't need RAM disk.

What on earth are you on about??

Did I ever say my SSD failed?

It's the only way I could make out what you were talking about to be a problem. But since it isn't a problem, you're just trying to make one out of whole cloth.

But if I continued going the same way it would.

And if you continue using your RAID 0 HDD's they will fail too. Truisms are not arguments either way.

Jesper

 

[SushiEater]

theswede wrote:

SushiEater wrote:

theswede wrote:

I set trim on, deadline scheduler, swapiness to 0 and use dynamic ram disk for temp files.

What is the point of having SSD if you use memory for the RAM disk?

Faster loading of files? Why is this a question?

Because most people use SSD to load files faster or have enough memory so they don't need RAM disk.

What on earth are you on about??

Did I ever say my SSD failed?

It's the only way I could make out what you were talking about to be a problem. But since it isn't a problem, you're just trying to make one out of whole cloth.

But if I continued going the same way it would.

And if you continue using your RAID 0 HDD's they will fail too. Truisms are not arguments either way.

Oh no, now I have 7 years left (predicted) before I had 8 years on the SSD. If I did the same thing (and I do quite often) to the Raid 0 absolutely nothing would have happen. 2 weeks of heavy use does not destroy HDDs. 2 weeks of heavy use destroyed 15% of SSD.

 

[theswede #100093]

And if you continue using your RAID 0 HDD's they will fail too. Truisms are not arguments either way.

Oh no, now I have 7 years left (predicted)

Which says nothing at all, because all it does is guess.

before I had 8 years on the SSD.

Which does not take into account how that particular drive handles internal management of wear and tear, reallocation of cells or any internal management at all. It's only a read and extrapolation from MWI, which is a non-controlled estimate to begin with. SSDlife is for the general case a waste of time as it uses uncontrolled estimates based on uncontrolled marketing influenced estimates to try to give a hard number.

If I did the same thing (and I do quite often) to the Raid 0 absolutely nothing would have happen.

If the MTBF of a HDD is 30 years and you have 30 drives, one will fail within a year.

2 weeks of heavy use does not destroy HDDs. 2 weeks of heavy use destroyed 15% of SSD.

Except it destroyed nothing at all since you have no failure. You're making a mountain out of a molehill while claiming the mountain ranges which exist are nothing to worry about.

Jesper

 

[kelpdiver]

theswede wrote:
SSDlife is for the general case a waste of time as it uses uncontrolled estimates based on uncontrolled marketing influenced estimates to try to give a hard number.

Worse - you have to pay money for it. Perfectly legit scam, taking advantage of latent fears of the SSD endurance boogeyman.

 

[SushiEater]

theswede wrote:

And if you continue using your RAID 0 HDD's they will fail too. Truisms are not arguments either way.

Oh no, now I have 7 years left (predicted)

Which says nothing at all, because all it does is guess.

Excuses, excuses. Come up with the real answer next time.

before I had 8 years on the SSD.

Which does not take into account how that particular drive handles internal management of wear and tear, reallocation of cells or any internal management at all. It's only a read and extrapolation from MWI, which is a non-controlled estimate to begin with. SSDlife is for the general case a waste of time as it uses uncontrolled estimates based on uncontrolled marketing influenced estimates to try to give a hard number.

I just got email from them after I showed them this thread. They said not to argue with you because it is a waste of time. Apparently you know more than they do about their software.

If I did the same thing (and I do quite often) to the Raid 0 absolutely nothing would have happen.

If the MTBF of a HDD is 30 years and you have 30 drives, one will fail within a year.

And you that how? Is there a writing on the wall?

2 weeks of heavy use does not destroy HDDs. 2 weeks of heavy use destroyed 15% of SSD.

Except it destroyed nothing at all since you have no failure. You're making a mountain out of a molehill while claiming the mountain ranges which exist are nothing to worry about.

Jesper

So you would rather want me to continue using it this way until failure? I make you a deal, send me brand new 120GB SSD plus $1000 for electricity use and wear and tear and my time and I will use my SSD to the end.