[kwlug-disc] SMART for SSDs

[B.S.]

* Aside from Khalid's specific case ...

Is it fair to say that, in general, an SSD will be OK ... until it's not. And even then the written data will still be readable / recoverable? (Unlike spinning drives where when they stop spinning, that's it / that's all.)

And ... I guess, aside from the filesystem suddenly going RO, how does a failing / failed SSD typically present itself?


* IIRC, threads involving Khalid and his / this specific SSD and his unhappiness with this particular defective model noted a spectacularly bad product that entirely fails (even reading) far sooner than reasonable. IIUC the issues with this model was substantially a one off for SSDs, not expected to appear in others. Particularly nothing more recent? (It's just too bad Khalid happened to be one of the unlucky victims. There was great unhappiness in the land ...)

>________________________________
> From: L.D. Paniak <ldpaniak at fourpisolutions.com>
>To: KWLUG discussion <kwlug-disc at kwlug.org> 
>Sent: Wednesday, July 29, 2015 1:37 PM
>Subject: Re: [kwlug-disc] SMART for SSDs
>
>

.
.
.
>I doubt you have anything to worry about here.
.
.
.
> I would expect to get 100TB of writes on the drive before starting
   to worry.
>The drive is rated for 300TB write.
>
>Tests of this kind of drive have gotten over 2.4PB of writes in
   before the drive quit:
>http://techreport.com/review/27909/the-ssd-endurance-experiment-theyre-all-dead
>
>
>On 07/29/2015 12:52 PM, Khalid Baheyeldin wrote:
>
>So, I had an SSD lying around for a couple of years (long story, bought to benchmark against spinning rust disks for a client project). My laptop disk died 1.5 months ago, and when checking SMART values for it, it was not good.

.
.
.