continuing disk crash issue

[Ian Bruntlett]

Hi Owen,

On Mon, 30 Jun 2025 at 15:18, Owen Thomas <owen.paul.thomas at gmail.com>
wrote:

> A few general observations I would like to make are that I recently
> purchased a set of three 32G USB sticks for backups (only needed one, but
> they came in sets of three), and found that each of them lasted for about
> two weeks before I couldn't read them any more. I thought: yea, they
> probably came from a bad batch. I talked to someone about this and he
> recommended that I get some spinning rust because of its reliability and
> size, However, being that 32G was more than enough for my indulgence, I'm
> not sure that I know what to do with something like 16T, and so I couldn't
> justify the cost.
>
> So, what I read here is more confirmation about this phenomenon although I
> more recently (about two months ago) seem to have found a solution to my
> problem by purchasing a 128G stick and I haven't experienced any problems
> so far. I noticed that the 32G sticks were not metallic, while the 128G one
> was. Perhaps (just a stab) this might be a reason for the reliability
> problems?
>
> Anyway these observations might provide further stimulation to this
> discussion and that might inform me yet again.
>

My experiences of USB flash sticks is that they still work after a fair
amount of time not being used. However, that may change :)

What make and model were your 32GB drives? Did you test them (e.g. with
something like f3) before you used them?

I tend to use SanDisk or Kingsto.

I received a couple of Lexar memory sticks from a friend this year. f3
reported some errors on one of them.here are my notes
I used the f3 package to test the first memory stick. I created a new
partition table, created one partition, formatted it as ext4. Ran f3write -
no errors. Ran f3read, to check the results of f3write and got back an
error stating that there were 29 faulty sectors. So I used mkfs.ext4 - in
particular using the parameter -cc to do a set of slow read-write tests of
the drive - had to abort it during the 2nd pass due to lack of time.
However, I believe that when the memory stick realises that a sector is
faulty, it swaps in a spare sector and goes its merry way. So I think the
drive is usable now.

The second drive went through the above test - new MBR, new partition
(ext4), told the partition not to reserve blocks for the superuser
(tune2fs) and a full f3write. f3read was happy with that. When I have time,
I'll be running mkfs.ext4 on it, to be sure :)


HTH,


Ian

-- 
-- ACCU - Professionalism in programming - http://www.accu.org
-- My writing - https://sites.google.com/site/ianbruntlett/
-- Free Software page -
https://github.com/ian-bruntlett/TECH-Manuals/blob/main/tm-free-software.md
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20250630/c0f5ae7c/attachment.html>