[Discuss] Failing WD Disks

[kentborg at borg.org (Kent Borg)]

On 5/18/23 12:21, Bill Ricker wrote:
> Indeed.
>
>     P.S. At least /dev/urandom, at least on my current machine, is a lot
>     faster than it used to be.
>
>
> Faster random may be less random ?

I do not think the quality of Linux's /dev/urandom has gone down; it got 
a major rewrite that I think was motivated to make the code better and 
cleaner, and the fact that it is also faster was a secondary bit. 
(Pretty sure /dev/random and /dev/urandom are now exactly the same.)


> Whether crypto-secure random is required for cloaking the data later 
> written on the disk depends upon from whom you're hiding!

It could *maybe* be useful from some traffic analysis type snooping. 
Which is why I quit doing it on new disks.


> And it's still better than having 0000 or DEADBEEF blocks as unused 
> space.)

I am a big fan of using decent random data for things like this because 
I don't trust storage systems to not be clever on me. Maybe I say to 
write a bunch of 0s and maybe it makes a note "a bunch of zeros here" 
and very quickly says yes. When I ask for that data it checks its notes 
and very quickly says 0s. Maybe it says "let me compress this data" in a 
more general-purpose way, which means I don't even want to write the 
same quality random data repeatedly.


As I think I have mentioned before, after I unmount one of these disks, 
if I leave it plugged in I can feel disk activity going on, for quite a 
while. I assume it is checking and remapping and who knows what. It 
might be doing maintaining access statistics and doing some caching 
something something tricks. It might be compressing. Using non-terrible 
random data is a way to make sure keep some of this from tricking me.


-kb