Force ssd failure?

[kkeeiiggaann]

i am interested in forcing the symptoms of ssd failure. ssds have enjoyed a controversial reputation over the past few years for failing due to wear, firmware problems, etc.

i'm sure that writing to a drive constantly for a long while would produce the results i'm looking for, but i'm curious about other methods of forcing a drive to fail. note that i am not interested in destroying a drive to wipe sensitive data. i'm merely interested in reproducing the types of behaviors a drive exhibits when it is "bricked," or failing beyond simple repair.

thoughts?

 

[Pinhedd]

Hit it with a hammer

Guaranteed to cause failure

 

[kkeeiiggaann]

i suppose i should have mentioned that i'm not looking to manually alter or destroy the drive.

 

[Pinhedd]

Much of the short-lifespan reputation is undeserved. SSDs will last 5-7 years under even heavy loads. The normal warranty period is 3 years.

 

[Pyree]

Flash it with the wrong firmware. Will take some effort to fix it.

Why do you need to brick a drive?

 

[vrumor]

Had a total of 6 SSDs over the past 2.5 years all in a RAID 0 environment and none have failed, hiccuped or otherwise poopooed on itself.

 

[coper210]

running SSD speed tests can take a tole on your SSD, probably would be the best way to brick it

 

[kkeeiiggaann]

@Pin - agreed, the reputation doesn't really add up. i guess it exists because people are very used to hdd reliability. hdds of course fail, but i think that consumer ssds earned the reputation when they had malfunctions that hdds don't, such as controller/firmware issues due to technology being pushed out a little sooner than it should have been. additionally, certain brands have a reputation for high DOA numbers and rate of failure due to (presumably) b-stock nand or sketchy firmware.
@Pyree - flashing with correct firmware would resolve the issue, and leave the controller and nand healthy...isn't that true? i don't really *need* to brick a drive. i *want* to. it isn't a matter of destroying sensitive data or anything.
@vrumor - i've heard that from a lot of ssd owners.
@coper - i suppose that amounts to writing to the drive repeatedly.

i've considered running tools with multi-pass drive-wiping overwrite algorithms designed for hdds, such as with dban. gutmann, perhaps. i wonder if there is a more efficient method of wearing out nand? i've seen defrag maligned thoroughly for the wear it causes on an ssd...is this justified, or is defrag merely unnecessary light wear?

other thoughts for pushing any part of the hardware in an ssd to the breaking point?

 

[Pyree]

I am wondering what will be the damage if you pull the power when you are doing a firmware upgrade.

Also try force more voltage down instead the normal 5v (mode the power cable so red 5v connect to a 12v power source, or higher voltage, love to see you try 120/240v or higher 😀. Be sure to make a video when you do 120/240v or higher). Burn out the circuit will definitely brick it beyond repairable.

Spoiler

Disclaimer: I can not be held responsible for any injury, death or property damage or total destruction.

 

[kkeeiiggaann]

well, that would do. what about bringing upon the kind of failure that people typically talk about when they talk about ssds...like death by wear?

 

[Pyree]

If you want to kill it by wearing the NAND, then you have to rewrite it many time. There is no other way, making your statement:

"i'm merely interested in reproducing the types of behaviors a drive exhibits when it is "bricked," or failing beyond simple repair. "

totally no point because there are no alternative way.

 

[kkeeiiggaann]

i guess the point would be, i'm looking for ideas on how to do so efficiently.

 

[vrumor]

I read this over a few times over the last week since its been up. My only questions is why do YOU want to test the reliability of an SSD. I cant even imagine how long it would take for it to force a failiure simply by rewriting info to it over and over. Seems like its fruitless. And would tie up your PC for a silly demonstration when the info is out there already.

 

[Rusting In Peace]

An interesting thread.

I'd just try and find out what the manufacturers do to stress test / measure lifespan and do exactly that.

 

[kkeeiiggaann]

@ vrumor: not really looking to test the reliability of an ssd. having said that, there are probably some useful kernels of knowledge from those who have/do test for reliability.

@ rusting: good thinking. do you know of any resources offhand?

 

[Dereck47]

Some members at Xtreme Systems Forum buy SSDs and run software that continuously writes to the SSD until it fails.

Here's the link: http://www.xtremesystems.org/forums/showthread.php?271063-SSD-Write-Endurance-25nm-Vs-34nm