Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.os.linux.misc > #77673 > unrolled thread

Taming The Data Destroyer

Back to article view | Back to comp.os.linux.misc

Contents

  Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-17 02:31 +0000
    Re: Taming The Data Destroyer rbowman <bowman@montana.com> - 2025-11-17 04:18 +0000
      Re: Taming The Data Destroyer Marc Haber <mh+usenetspam1118@zugschl.us> - 2025-11-17 10:42 +0100
        Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 01:10 +0000
          Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-17 21:08 -0500
      Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-17 12:51 +0100
    Re: Taming The Data Destroyer jayjwa <jayjwa@atr2.ath.cx.invalid> - 2025-11-17 12:55 -0500
      Re: Taming The Data Destroyer not@telling.you.invalid (Computer Nerd Kev) - 2025-11-18 07:06 +1000
        Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-17 19:00 -0500
          Re: Taming The Data Destroyer Computer Nerd Kev <not@telling.you.invalid> - 2025-11-18 16:31 +1000
            Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-18 07:46 +0100
              Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-18 02:12 -0500
                Re: Taming The Data Destroyer rbowman <bowman@montana.com> - 2025-11-18 23:02 +0000
            Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-18 02:01 -0500
              Re: Taming The Data Destroyer Ralf Fassel <ralfixx@gmx.de> - 2025-11-18 16:24 +0100
                Re: Taming The Data Destroyer The Natural Philosopher <tnp@invalid.invalid> - 2025-11-18 15:30 +0000
                  Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 22:31 +0000
                Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-18 17:49 +0100
                  Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-19 18:32 -0500
                    Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-20 12:57 +0100
                      Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-20 22:53 +0000
                        Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-20 22:04 -0500
                      Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-20 22:17 -0500
            Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 07:18 +0000
              Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-18 02:34 -0500
                Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-18 08:46 +0100
                  Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 08:09 +0000
                    Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-18 17:55 +0100
                      Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 19:35 +0000
                        Re: Taming The Data Destroyer not@telling.you.invalid (Computer Nerd Kev) - 2025-11-19 06:48 +1000
                          Re: Taming The Data Destroyer The Natural Philosopher <tnp@invalid.invalid> - 2025-11-18 21:40 +0000
                            Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 22:27 +0000
                          Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 22:26 +0000
                          Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-19 18:23 -0500
                        Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-19 02:41 +0100
                  Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-18 03:17 -0500
                    Re: Taming The Data Destroyer The Natural Philosopher <tnp@invalid.invalid> - 2025-11-18 12:12 +0000
                      Re: Taming The Data Destroyer "Carlos E.R." <robin_listas@es.invalid> - 2025-11-18 17:53 +0100
                        Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 19:32 +0000
              Re: Taming The Data Destroyer Anssi Saari <anssi.saari@usenet.mail.kapsi.fi> - 2025-11-18 14:46 +0200
        Re: Taming The Data Destroyer candycanearter07 <candycanearter07@candycanearter07.nomail.afraid> - 2025-11-20 18:40 +0000
      Re: Taming The Data Destroyer Lawrence D’Oliveiro <ldo@nz.invalid> - 2025-11-18 00:58 +0000
        Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-17 21:04 -0500
          Re: Taming The Data Destroyer rbowman <bowman@montana.com> - 2025-11-18 06:32 +0000
            Re: Taming The Data Destroyer c186282 <c186282@nnada.net> - 2025-11-18 02:07 -0500

Page 1 of 3  [1] 2 3  Next page →

#77673 — Taming The Data Destroyer

I use the “dd” command fairly frequently, to put new versions of
SystemRescue on the USB stick I dedicate to that purpose, for example.
The command has been nicknamed the “data destroyer”, because while
doing sector-level copies as root, it can be all too easy to write to
the wrong device, and destroy an important filesystem.

The “lsblk” command can be handy for confirming which block device you
should be writing to, before pressing Enter on that fateful command.
There’s lots of information it can display, to help you identify the
right device. For example, a command like

    lsblk -o NAME,SIZE,SERIAL,WWN,VENDOR,MODEL

shows a little tree display of devices and partitions in the first
column, followed by a lot of info you can use to identify what is
currently plugged in. I tried a couple of USB sticks, and they showed
amazingly long strings for their serial numbers. My SystemRescue stick
shows the vendor as “TOSHIBA” and the model as “TransMemory”.

Interestingly, both Western Digital and Seagate drives show the vendor
as “ATA”, though the model name begins with “WDC” for the one and “ST”
for the other. The SSD boot drives in both my main workstation and
laptop display blank for the vendor, but again the model field shows
some vendor-specific information.

[toc] | [next] | [standalone]

#77678

On Mon, 17 Nov 2025 02:31:41 -0000 (UTC), Lawrence D’Oliveiro wrote:

>     lsblk -o NAME,SIZE,SERIAL,WWN,VENDOR,MODEL

On Ubuntu  'lsblk -e7' is your friend.

[toc] | [prev] | [next] | [standalone]

#77681

rbowman <bowman@montana.com> wrote:
>On Mon, 17 Nov 2025 02:31:41 -0000 (UTC), Lawrence D’Oliveiro wrote:
>
>>     lsblk -o NAME,SIZE,SERIAL,WWN,VENDOR,MODEL
>
>On Ubuntu  'lsblk -e7' is your friend.

You can also use the nice links from /dev/disk in the dd command to
avoid hitting the wrong device.

Greetings
Marc
-- 
----------------------------------------------------------------------------
Marc Haber         |   " Questions are the         | Mailadresse im Header
Rhein-Neckar, DE   |     Beginning of Wisdom "     | 
Nordisch by Nature | Lt. Worf, TNG "Rightful Heir" | Fon: *49 6224 1600402

[toc] | [prev] | [next] | [standalone]

#77697

On Mon, 17 Nov 2025 10:42:06 +0100, Marc Haber wrote:

> You can also use the nice links from /dev/disk in the dd command to
> avoid hitting the wrong device.

WWN links can be useful on storage servers, but USB sticks don’t have 
WWNs. /dev/disk/by-id does contain entries for them, in some strange 
format.

[toc] | [prev] | [next] | [standalone]

#77699

On 11/17/25 20:10, Lawrence D’Oliveiro wrote:
> On Mon, 17 Nov 2025 10:42:06 +0100, Marc Haber wrote:
> 
>> You can also use the nice links from /dev/disk in the dd command to
>> avoid hitting the wrong device.
> 
> WWN links can be useful on storage servers, but USB sticks don’t have
> WWNs. /dev/disk/by-id does contain entries for them, in some strange
> format.

   Again, a smart reason to use the good old utilities
   that decode all that mess FOR you.

   In Python, I sometimes load-up /proc/mounts because
   then, without further BS, you can do a quick string
   search to see if something you THINK is mounted IS
   actually mounted. Sometimes they aren't. Always
   included this check in backup apps.

[toc] | [prev] | [next] | [standalone]

#77683

On 2025-11-17 05:18, rbowman wrote:
> On Mon, 17 Nov 2025 02:31:41 -0000 (UTC), Lawrence D’Oliveiro wrote:
> 
>>      lsblk -o NAME,SIZE,SERIAL,WWN,VENDOR,MODEL
> 
> On Ubuntu  'lsblk -e7' is your friend.

        -e, --exclude list
            Exclude the devices specified by the
            comma-separated list of major device
            numbers. Note that RAM disks (major=1) are
            excluded by default if --all is not
            specified. The filter is applied to the
            top-level devices only. This may be
            confusing for --list output format where
            hierarchy of the devices is not obvious.

Major 7 are "loop" devices.


My favourite is:

lsblk --output NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,VENDOR,MODEL,SERIAL,REV,ZONED,ALIGNMENT

-- 
Cheers, Carlos.
ES🇪🇸, EU🇪🇺;

[toc] | [prev] | [next] | [standalone]

#77689

Lawrence D’Oliveiro <ldo@nz.invalid> writes:

> The command has been nicknamed the “data destroyer”, because while
> doing sector-level copies as root, it can be all too easy to write to
> the wrong device, and destroy an important filesystem.
This is likely because new users come to Linux and don't know what they
are doing and end up just running commands they see online that might
not pertain to their system's setup. This, or Windows users write to the
wrong device, not being familiar with Linux block devices. "rm -rf" (or
just "rm") can be pointed in the wrong place just as easy when run as root.

There's a Windows IRC channel that says something to the effect of
"don't use dd to write images!" in the channel topic.

> The “lsblk” command can be handy for confirming which block device you
> should be writing to
What, you don't parse /proc/partitions yourself?

> Interestingly, both Western Digital and Seagate drives show the vendor
> as “ATA”
I see that here with my WDC, too. Odd. The USB stick shows "Memorex" as
it should. 

-- 
PGP Key ID: 781C A3E2 C6ED 70A6 B356  7AF5 B510 542E D460 5CAE
       "The Internet should always be the Wild West!"

[toc] | [prev] | [next] | [standalone]

#77692

jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>> The "lsblk" command can be handy for confirming which block device you
>> should be writing to
> What, you don't parse /proc/partitions yourself?

I like "fdisk -l". The "lsblk" output is more condensed, but that
also makes it easier to get mixed up between lines. On a system
with lots of drives/partitions the "lsblk" output is obviously
better, but I'd double check with "fdisk -l /dev/[drive]" before
doing something as serious as writing over a partition's contents.

-- 
__          __
#_ < |\| |< _#

[toc] | [prev] | [next] | [standalone]

#77693

On 11/17/25 16:06, Computer Nerd Kev wrote:
> jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
>> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>>> The "lsblk" command can be handy for confirming which block device you
>>> should be writing to
>> What, you don't parse /proc/partitions yourself?
> 
> I like "fdisk -l". The "lsblk" output is more condensed, but that
> also makes it easier to get mixed up between lines. On a system
> with lots of drives/partitions the "lsblk" output is obviously
> better, but I'd double check with "fdisk -l /dev/[drive]" before
> doing something as serious as writing over a partition's contents.

   Linux has several commands that can do total
   damage instantly. Wrong use of 'fdisk' is one
   of those and 'dd' is oft nicknamed "disk
   destroyer" most often because people get
   source and destination reversed. Then there
   is always the recursive param for 'rm' :-)

   You can do such damage with DOS/Win too ...
   nothing especially evil about Linux/UNIX.
   Such lower-level commands are NECESSARY for
   managing any traditional sort of computer.

   As for "taming" ... that's YOUR job.

   And yea, always start with 'lsblk' ... will usually
   tell you everything you need. Is that disk/drive
   really connected ? Where is it, or is it, mounted ?

[toc] | [prev] | [next] | [standalone]

#77703

c186282 <c186282@nnada.net> wrote:
> On 11/17/25 16:06, Computer Nerd Kev wrote:
>> jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
>>> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>>>> The "lsblk" command can be handy for confirming which block device you
>>>> should be writing to
>>> What, you don't parse /proc/partitions yourself?
>> 
>> I like "fdisk -l". The "lsblk" output is more condensed, but that
>> also makes it easier to get mixed up between lines. On a system
>> with lots of drives/partitions the "lsblk" output is obviously
>> better, but I'd double check with "fdisk -l /dev/[drive]" before
>> doing something as serious as writing over a partition's contents.
> 
>   Linux has several commands that can do total
>   damage instantly. Wrong use of 'fdisk' is one
>   of those

No fdisk won't do damage instantly when you just run one command
from the command line, modifying partitions is done in its
interactive mode. You can't stuff things up just by attempting an
"fdisk -l".

-- 
__          __
#_ < |\| |< _#

[toc] | [prev] | [next] | [standalone]

#77705

On 2025-11-18 07:31, Computer Nerd Kev wrote:
> c186282 <c186282@nnada.net> wrote:
>> On 11/17/25 16:06, Computer Nerd Kev wrote:
>>> jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
>>>> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>>>>> The "lsblk" command can be handy for confirming which block device you
>>>>> should be writing to
>>>> What, you don't parse /proc/partitions yourself?
>>>
>>> I like "fdisk -l". The "lsblk" output is more condensed, but that
>>> also makes it easier to get mixed up between lines. On a system
>>> with lots of drives/partitions the "lsblk" output is obviously
>>> better, but I'd double check with "fdisk -l /dev/[drive]" before
>>> doing something as serious as writing over a partition's contents.
>>
>>    Linux has several commands that can do total
>>    damage instantly. Wrong use of 'fdisk' is one
>>    of those
> 
> No fdisk won't do damage instantly when you just run one command
> from the command line, modifying partitions is done in its
> interactive mode. You can't stuff things up just by attempting an
> "fdisk -l".

You can write partition tables from the command line. I have never 
tried, but it is possible. Some partitioning software can save a backup 
file of the table, and can restore the table from that file.

That said, writing a wrong partition table is not destructive, it can be 
undone. Instantly if you have a backup, or after a long scan of the 
entire disk if you don't. These tools normally do not format the 
partitions (fdisk doesn't), or erase data, but there are tools that 
combine a format phase.

-- 
Cheers, Carlos.
ES🇪🇸, EU🇪🇺;

[toc] | [prev] | [next] | [standalone]

#77708

On 11/18/25 01:46, Carlos E.R. wrote:
> On 2025-11-18 07:31, Computer Nerd Kev wrote:
>> c186282 <c186282@nnada.net> wrote:
>>> On 11/17/25 16:06, Computer Nerd Kev wrote:
>>>> jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
>>>>> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>>>>>> The "lsblk" command can be handy for confirming which block device 
>>>>>> you
>>>>>> should be writing to
>>>>> What, you don't parse /proc/partitions yourself?
>>>>
>>>> I like "fdisk -l". The "lsblk" output is more condensed, but that
>>>> also makes it easier to get mixed up between lines. On a system
>>>> with lots of drives/partitions the "lsblk" output is obviously
>>>> better, but I'd double check with "fdisk -l /dev/[drive]" before
>>>> doing something as serious as writing over a partition's contents.
>>>
>>>    Linux has several commands that can do total
>>>    damage instantly. Wrong use of 'fdisk' is one
>>>    of those
>>
>> No fdisk won't do damage instantly when you just run one command
>> from the command line, modifying partitions is done in its
>> interactive mode. You can't stuff things up just by attempting an
>> "fdisk -l".
> 
> You can write partition tables from the command line. I have never 
> tried, but it is possible. Some partitioning software can save a backup 
> file of the table, and can restore the table from that file.
> 
> That said, writing a wrong partition table is not destructive, it can be 
> undone. Instantly if you have a backup, or after a long scan of the 
> entire disk if you don't. These tools normally do not format the 
> partitions (fdisk doesn't), or erase data, but there are tools that 
> combine a format phase.

   A problem with newbies/oblivious is that they tend
   to search the web for fixes to whatever minor prob
   and then just cut-n-paste. They don't/can't even
   consider what those 13 params mean.

   Do they "deserve it" ? That is one way of looking
   at it ... but too often it's the sites offering
   the canned fixes that fail to explain clearly.

[toc] | [prev] | [next] | [standalone]

#77736

On Tue, 18 Nov 2025 02:12:22 -0500, c186282 wrote:

>    A problem with newbies/oblivious is that they tend to search the web
>    for fixes to whatever minor prob and then just cut-n-paste. They
>    don't/can't even consider what those 13 params mean.

Color me oblivious. Most of the things I've done with PowerShell have been 
a blind cut'n'paste hoping for the best. At this point in life learning 
PowerShell is Priority 7.  (7 was the priority we assigned to requests 
that meant 'nobody is ever going to be bored enough to touch that dark 
hole'.)

I sometimes see people complaining about a bug that hasn't been resolved 
in years. That usually means one or more people gently prodded it, sniffed 
the air, and ran like hell. 

[toc] | [prev] | [next] | [standalone]

#77706

On 11/18/25 01:31, Computer Nerd Kev wrote:
> c186282 <c186282@nnada.net> wrote:
>> On 11/17/25 16:06, Computer Nerd Kev wrote:
>>> jayjwa <jayjwa@atr2.ath.cx.invalid> wrote:
>>>> Lawrence D'Oliveiro <ldo@nz.invalid> writes:
>>>>> The "lsblk" command can be handy for confirming which block device you
>>>>> should be writing to
>>>> What, you don't parse /proc/partitions yourself?
>>>
>>> I like "fdisk -l". The "lsblk" output is more condensed, but that
>>> also makes it easier to get mixed up between lines. On a system
>>> with lots of drives/partitions the "lsblk" output is obviously
>>> better, but I'd double check with "fdisk -l /dev/[drive]" before
>>> doing something as serious as writing over a partition's contents.
>>
>>    Linux has several commands that can do total
>>    damage instantly. Wrong use of 'fdisk' is one
>>    of those
> 
> No fdisk won't do damage instantly when you just run one command
> from the command line, modifying partitions is done in its
> interactive mode. You can't stuff things up just by attempting an
> "fdisk -l".

   Not with '-l' ... but if you supply MORE params ....

   Few Linux CL utils DEMAND interactive mode, there
   are work-arounds ... which MIGHT be found on
   'helpful' web sites ... intended for fully
   automatic purposes.

   In any case, newbies need to be CAREFUL - don't
   just cut-n-paste anything you find. If you're
   not sure - DON'T !

[toc] | [prev] | [next] | [standalone]

#77721

* c186282 <c186282@nnada.net>
|   In any case, newbies need to be CAREFUL - don't
|   just cut-n-paste anything you find. If you're
|   not sure - DON'T !

But... but... the AI said it was ok to use that command! ;-)

Your advice holds not only for newbies.  When I search for the solution
on some "non-trivial" problem, sometimes I wonder what the magic
incantation on some search result is - a new clever way to achieve the
result, monkey-see-monkey-do copied from somewhere else, or simply
nonsense (possibly in combination with #2).  More often than not it
turns out to be the latter...

R'

[toc] | [prev] | [next] | [standalone]

#77722

On 18/11/2025 15:24, Ralf Fassel wrote:
> * c186282 <c186282@nnada.net>
> |   In any case, newbies need to be CAREFUL - don't
> |   just cut-n-paste anything you find. If you're
> |   not sure - DON'T !
> 
> But... but... the AI said it was ok to use that command! ;-)
> 
> Your advice holds not only for newbies.  When I search for the solution
> on some "non-trivial" problem, sometimes I wonder what the magic
> incantation on some search result is - a new clever way to achieve the
> result, monkey-see-monkey-do copied from somewhere else, or simply
> nonsense (possibly in combination with #2).  More often than not it
> turns out to be the latter...
> 
> R'
My biggest moan is the constantly moving target of code being 'improved' 
by bright eyed bushy tailed compScis to ensure that nothing you find on 
the net is in fact applicable to the version you are now running.

Systemd has made this far far worse.

-- 
"Strange as it seems, no amount of learning can cure stupidity, and 
higher education positively fortifies it."

    - Stephen Vizinczey

[toc] | [prev] | [next] | [standalone]

#77735

On Tue, 18 Nov 2025 15:30:42 +0000, The Natural Philosopher wrote:

> My biggest moan is the constantly moving target of code being
> 'improved' by bright eyed bushy tailed compScis to ensure that
> nothing you find on the net is in fact applicable to the version you
> are now running.

You are free to stick with old versions. That’s part of what “Free
software” means.

> Systemd has made this far far worse.

systemd has an official site where you will find authoritative
reference documentation <https://systemd.io/>.

Also links to this list
<https://0pointer.de/blog/projects/the-biggest-myths.html> of the
worst of the same old tired myths that misinformed people insist on
continually repeating about systemd ...

[toc] | [prev] | [next] | [standalone]

#77724

On 2025-11-18 16:24, Ralf Fassel wrote:
> * c186282 <c186282@nnada.net>
> |   In any case, newbies need to be CAREFUL - don't
> |   just cut-n-paste anything you find. If you're
> |   not sure - DON'T !
> 
> But... but... the AI said it was ok to use that command! ;-)
> 
> Your advice holds not only for newbies.  When I search for the solution
> on some "non-trivial" problem, sometimes I wonder what the magic
> incantation on some search result is - a new clever way to achieve the
> result, monkey-see-monkey-do copied from somewhere else, or simply
> nonsense (possibly in combination with #2).  More often than not it
> turns out to be the latter...
> 
> R'

I have found ChatGpt very useful to help me with solving problems with 
my 8 disk software raid 6 array. It does diagnose the problem correctly 
and generates the correct command line concoctions. Of course I check 
the lines, but they look correct, and do work.

-- 
Cheers, Carlos.
ES🇪🇸, EU🇪🇺;

[toc] | [prev] | [next] | [standalone]

#77748

On 11/18/25 11:49, Carlos E.R. wrote:
> On 2025-11-18 16:24, Ralf Fassel wrote:
>> * c186282 <c186282@nnada.net>
>> |   In any case, newbies need to be CAREFUL - don't
>> |   just cut-n-paste anything you find. If you're
>> |   not sure - DON'T !
>>
>> But... but... the AI said it was ok to use that command! ;-)
>>
>> Your advice holds not only for newbies.  When I search for the solution
>> on some "non-trivial" problem, sometimes I wonder what the magic
>> incantation on some search result is - a new clever way to achieve the
>> result, monkey-see-monkey-do copied from somewhere else, or simply
>> nonsense (possibly in combination with #2).  More often than not it
>> turns out to be the latter...
>>
>> R'
> 
> I have found ChatGpt very useful to help me with solving problems with 
> my 8 disk software raid 6 array. It does diagnose the problem correctly 
> and generates the correct command line concoctions. Of course I check 
> the lines, but they look correct, and do work.

   Until one melts your drive  :-)

   Of late I'd been doing a lot of stuff with 'ffmpeg'.
   It's an all-purpose tool BUT to get what YOU want
   often involves ten or fifteen CL flags and such
   AND in a particular order. Plenty of 'examples' to
   be found online - not all agree with each other.

   Basically you need the skills to put the evil eye on
   such 'free' code BEFORE you run it, looking for that
   little 'gotcha'. Look up those commands, their params,
   the 'why', before you go too far. Being a computer jock,
   even in the AI age, is a learned skill ... not something
   Granny can just jump into.

[toc] | [prev] | [next] | [standalone]

#77754

On 2025-11-20 00:32, c186282 wrote:
> On 11/18/25 11:49, Carlos E.R. wrote:
>> On 2025-11-18 16:24, Ralf Fassel wrote:
>>> * c186282 <c186282@nnada.net>
>>> |   In any case, newbies need to be CAREFUL - don't
>>> |   just cut-n-paste anything you find. If you're
>>> |   not sure - DON'T !
>>>
>>> But... but... the AI said it was ok to use that command! ;-)
>>>
>>> Your advice holds not only for newbies.  When I search for the solution
>>> on some "non-trivial" problem, sometimes I wonder what the magic
>>> incantation on some search result is - a new clever way to achieve the
>>> result, monkey-see-monkey-do copied from somewhere else, or simply
>>> nonsense (possibly in combination with #2).  More often than not it
>>> turns out to be the latter...
>>>
>>> R'
>>
>> I have found ChatGpt very useful to help me with solving problems with 
>> my 8 disk software raid 6 array. It does diagnose the problem 
>> correctly and generates the correct command line concoctions. Of 
>> course I check the lines, but they look correct, and do work.
> 
>    Until one melts your drive  :-)
> 
>    Of late I'd been doing a lot of stuff with 'ffmpeg'.
>    It's an all-purpose tool BUT to get what YOU want
>    often involves ten or fifteen CL flags and such
>    AND in a particular order. Plenty of 'examples' to
>    be found online - not all agree with each other.
> 
>    Basically you need the skills to put the evil eye on
>    such 'free' code BEFORE you run it, looking for that
>    little 'gotcha'. Look up those commands, their params,
>    the 'why', before you go too far. Being a computer jock,
>    even in the AI age, is a learned skill ... not something
>    Granny can just jump into.

And if you read the entire man page and study it, for version 3, say, 
then comes another version which has changes and some of the 
combinations you used now do not work or have changed. They are now at 
version 8.

I suppose that's the reason that openSUSE ships all versions from 3 onwards.


ChatGpt will happily try to get you a CLI that produces what you work. 
But it also fails, and after several cycles you get mad.

-- 
Cheers, Carlos.
ES🇪🇸, EU🇪🇺;

[toc] | [prev] | [next] | [standalone]

Page 1 of 3  [1] 2 3  Next page →

Back to top | Article view | comp.os.linux.misc

csiph-web