Re: HDD vs SSD

John McGaw wrote:
> On 5/29/2020 3:42 PM, Paul wrote:
>> The SSD isn't in the mood for that sort of thing, and will
>> go "kerplunk" just like that. Turn it on some morning, and
>> "inaccessible boot volume" or equivalent.
>>
>> With the hard drive, a lazy user could "make a backup just
>> before it dies". Whereas with an SSD, you really have to
>> plan ahead. An SSD is a Ninja, in life and in death.
> 
> Dunno. I think that for HDD fails over the years I've had as many simply 
> croak without warning as gave some notice beforehand. Fortunately (?) 
> this has made me super paranoid about backups so I have multiple levels 
> and locations of coverage when the inevitable happens yet again.
> 
> Thinking back I'm pretty sure that I've had only one real SSD fail but 
> at the moment the brand doesn't register (OCZ?) -- just that it was a 
> SATA device and was back when 120gB (or what ever it was) was just 
> getting down to real-world pricing. I'm just using Samsung drives for 
> internal storage now but I do have three odd cheapish Pioneer drives 
> https://www.amazon.com/gp/product/B07H8FN3ZV for backup in external 
> USB-3 enclosures that are kept in a media safe in the office. Looking 
> back it seems that my first SSD was an Intel 80gB bought almost exactly 
> ten years back -- made me feel like the king of the world the fist time 
> I booted from it. :-)
> 

The Intel ones in particular, must be watched.

After 3000 write cycles to the Intel 545S, the Intel drive
policy is to "brick the unit", neither accepting read nor write
operations.

It's because of this abject "functionality policy" (which can be
monitored if you religiously watch your SMART output), I have
to describe the drives as capable of "Ninja failure".

While they can have firmware failures (like early OCZ), it's
also possible to suffer the effects of the "bricking policy"
at end of life. Some brands brick read-only, some just... brick.

End of life policies:

    1) There is a brand which just keeps running past 3000.
       Maybe you happen to catch signs of various errors, as a hint
       you really should look at the SMART. This isn't a particularly
       good policy, but it does leave it to your own initiative to
       "manage" your drives. Some drives manage to go well past
       their datasheet TBW rating (like do writes to them for another year).

    2) There is a brand which goes read-only. This is clever, because
       it allows doing a backup, and because Windows *hates* read-only
       behaviors, it takes no time at all to notice :-)

    3) Then there's the Intel behavior. Total brickage. Data locked up.
       No backup ? Tough Beanz.

While it's unlikely there will be brickage reports in a Newegg
review, you should check somewhere, for comments about what
the end-of-life policy is. It should be documented somewhere.

HTH,
     Paul