Undedr /0 (in addition to everything else that I don't want systems to
step on) I have canned mountpoints for every partition on every
imaginable device such as disks from /dev/sda to /dev/sdg. One such is
/0/sda4 which theoretically serves as a mount-POINT for dev/sda4. This
partition happens to be my Leap-15.3 partition and I want to mount it so
that I might copy one user's home folder over to my TW system (on
dev/sda5). I just created this user with a specific look-and-feel and
the LAST thing I want to do is repeat the exercise to set up under TW
and then 5 other distros as well.

This would/should go like

fdisk -l
...to confirm that sda IS the disk seen as primary, a step made necessary
ONLY by Leap which ALONE in all my distros continues to sometimes list
it as /dev/sdb.

There *is* this fdisk item

/dev/mapper/Samsung_SSD_860_EVO_1TB_S59VNJ0N419951T-part4

but I can't really use it because on one hand I can only guess that it
refers to /dev/sda4 and on the other typing yard-and-a-half long names
isn't my forte.

mount /dev/sda4 /0/sa04

FileManager copy /0/sa04/home/userX to /home/userX

make all /home/userX owned by userX

now for the real run:

# umount /dev/sda4
umount: /dev/sda4: not mounted.

# mount /dev/sda4 /0/sa04
mount: /0/sa04: /dev/sda4 already mounted or mount point busy.
dmesg(1) may have more information after failed mount system call.

# dmesg
[ 222.582351] /dev/sda4: Can't open blockdev
[ 313.515808] /dev/sda4: Can't open blockdev
[ 366.135462] /dev/sda4: Can't open blockdev
[ 416.200151] /dev/sda4: Can't open blockdev

If I click the USB icon in SystTray I can select ALL devices and after
an hour's worth of eyestrain I can make out /dev/sda4. One option is to
"Open in File Manager" (not even 'boo' as to what moint-POINT is being
talked about!). Why is this important? BECAUSE the above message said
"or mountpoint busy" which couldn't possibley refer to the attempted
mountpoint of /0/sa04 because the systemn doesn't know about my canned
mountpoints. It would be nice all the same to have some clue as to what
moount-POINT was being talked about.

So, how am I supposed to get to /dev/sda4 for the stated copy
without taking it out of the mobile tray and plugging it in as a USB
thing on a USB pigtail?

On 2023-03-31 13:14, bad sector wrote:
>
> Undedr /0 (in addition to everything else that I don't want systems to
> step on) I have canned mountpoints for every partition on every
> imaginable device such as disks from /dev/sda to /dev/sdg. One such is
> /0/sda4 which theoretically serves as a mount-POINT for dev/sda4. This
> partition happens to be my Leap-15.3 partition and I want to mount it so
> that I might copy one user's home folder over to my TW system (on
> dev/sda5). I just created this user with a specific look-and-feel and
> the LAST thing I want to do is repeat the exercise to set up under TW
> and then 5 other distros as well.
>
> This would/should go like
>
> fdisk -l
> ..to confirm that sda IS the disk seen as primary, a step made necessary
> ONLY by Leap which ALONE in all my distros continues to sometimes list
> it as /dev/sdb.

Plain simple:

DO NOT use device names such as /dev/sda. Use the new nomenclature, no
matter if you think they are too long.

Then, the command to see the information you need is:

lsblk --output
NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT
| less -S

(one single line). Keep it in your important_notes.txt and paste it when
needed; use as root.)

>
> There *is* this fdisk item
>
>   /dev/mapper/Samsung_SSD_860_EVO_1TB_S59VNJ0N419951T-part4
>
> but I can't really use it because on one hand I can only guess that it
> refers to /dev/sda4 and on the other typing yard-and-a-half long names
> isn't my forte.
>
>
> mount /dev/sda4 /0/sa04
>
> FileManager copy /0/sa04/home/userX to /home/userX
>
> make all /home/userX owned by userX
>
> now for the real run:
>
> # umount /dev/sda4
> umount: /dev/sda4: not mounted.
>
> # mount /dev/sda4 /0/sa04
> mount: /0/sa04: /dev/sda4 already mounted or mount point busy.
>        dmesg(1) may have more information after failed mount system call.
>
>
> # dmesg
> [  222.582351] /dev/sda4: Can't open blockdev
> [  313.515808] /dev/sda4: Can't open blockdev
> [  366.135462] /dev/sda4: Can't open blockdev
> [  416.200151] /dev/sda4: Can't open blockdev

Well, you have a hardware problem.

>
>
> If I click the USB icon in SystTray I can select ALL devices and after
> an hour's worth of eyestrain I can make out /dev/sda4. One option is to
> "Open in File Manager" (not even 'boo' as to what moint-POINT is being
> talked about!). Why is this important? BECAUSE the above message said
> "or mountpoint busy" which couldn't possibley refer to the attempted
> mountpoint of /0/sa04 because the systemn doesn't know about my canned
> mountpoints. It would be nice all the same to have some clue as to what
> moount-POINT was being talked about.

mountpoint is clearly is "/0/sa04". It says so.

>
> So, how am I supposed to get to /dev/sda4 for the stated copy
> without taking it out of the mobile tray and plugging it in as a USB
> thing on a USB pigtail?

Your post is confusing, and you use the wrong command to get the
information.

--
Cheers, Carlos.

On 3/31/23 07:41, Carlos E.R. wrote:
> lsblk --output
> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S

Cool, but it doesn't answer two of my questions :-)

1
In the attempted command "mount /dev/sda4 /0/sa04" the
'mountpoint' was /0/sa04. The error message stating that
THE mountpoint was in use is BS needing to be both clarified
AND fixed.

2
How do I mount the prinmary (or only) disk partition-4
under /0/sa04?

On 2023-03-31 21:03, bad sector wrote:
> On 3/31/23 07:41, Carlos E.R. wrote:
>> lsblk --output
>> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S
>
> Cool, but it doesn't answer two of my questions :-)
>
> 1
> In the attempted command "mount  /dev/sda4 /0/sa04" the
> 'mountpoint' was /0/sa04. The error message stating that
> THE mountpoint was in use is BS needing to be both clarified
> AND fixed.

I pose that pretense is just an error in the blob that sits between the
chair and the computer :-P :-D

If the message says the mount point is in use, then it is. Actually, the
message said two things with and 'or' in between.

You can use the command "mount" to see what is actually mounted and
where. And "lsof" to see what is busy.

>
> 2
> How do I mount the prinmary (or only) disk partition-4
> under /0/sa04?

You have to find first where is the error.

--
Cheers, Carlos.

On 3/31/23 16:27, Carlos E.R. wrote:
> On 2023-03-31 21:03, bad sector wrote:
>> On 3/31/23 07:41, Carlos E.R. wrote:
>>> lsblk --output
>>> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S
>>
>> Cool, but it doesn't answer two of my questions :-)
>>
>> 1
>> In the attempted command "mount  /dev/sda4 /0/sa04" the
>> 'mountpoint' was /0/sa04. The error message stating that
>> THE mountpoint was in use is BS needing to be both clarified
>> AND fixed.
>
> I pose that pretense is just an error in the blob that sits between the
> chair and the computer :-P  :-D
>
> If the message says the mount point is in use, then it is. Actually, the
> message said two things with and 'or' in between.
>
>
> You can use the command "mount" to see what is actually mounted and
> where. And "lsof" to see what is busy.
>
>
>>
>> 2
>> How do I mount the prinmary (or only) disk partition-4
>> under /0/sa04?
>
> You have to find first where is the error.
>

The first error I see is in the error message i.e. I can mount a device
under many mountpoints

# mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386 /0/sa04
# mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386 /0/sb04
# mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386 /0/sc04

never reports "already mounted", there may well be a limit of how many
mountpoints I can use but that's above my paygrade.

# umount /0/sc04

leaves only the first 2 mountpoints mounted under

# umount PARTUUID=04edd347-8460-c449-a350-4716a10d0386

unmounts it from the next to last remaining mountpoint

and so on until there are no more.

So in response to

mount /dev/sda4 /0/sa04

the message "already mounted OR mount point busy" can only mean that the
mountpoint (/0/sa04) is busy i.e. some device is already mounted under
/0/sa04. But /0/sa04 is *my* mountpoint, the system doesn't know it
exists so

1
how can it mount anything under it?

2
how does it get the authority to mount ANY partition without being
'commanded' by root to mount it either in fstab or explicitly?

lsof doesn't tell me anything significant

and your lsblok one-liner (thanks BTW) just ends with

"[14551.227219] /dev/sda4: Can't open blockdev"

Maybe THAT's the problem, "the criminal is using dev-names!!!!"

Well, I'nm NOT going to start copy-pasting two-yard long UUIDS just
becasue the NSA wanted absolute traceability on everything including La
Plume de Ma Tante. As an extreme worst case scenario (should I ever end
up with my back to the wall) I might either label partitions with the
last VERY-FEW digits of the disk RealHardware model number and partition
number, or I might just boot Slackware and do 'mount /dev/sda4 /0/sa04'.

On 2023-04-01 01:46, bad sector wrote:
> On 3/31/23 16:27, Carlos E.R. wrote:
>> On 2023-03-31 21:03, bad sector wrote:
>>> On 3/31/23 07:41, Carlos E.R. wrote:
>>>> lsblk --output
>>>> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S
>>>
>>> Cool, but it doesn't answer two of my questions :-)
>>>
>>> 1
>>> In the attempted command "mount  /dev/sda4 /0/sa04" the
>>> 'mountpoint' was /0/sa04. The error message stating that
>>> THE mountpoint was in use is BS needing to be both clarified
>>> AND fixed.
>>
>> I pose that pretense is just an error in the blob that sits between
>> the chair and the computer :-P  :-D
>>
>> If the message says the mount point is in use, then it is. Actually,
>> the message said two things with and 'or' in between.
>>
>>
>> You can use the command "mount" to see what is actually mounted and
>> where. And "lsof" to see what is busy.
>>
>>
>>>
>>> 2
>>> How do I mount the prinmary (or only) disk partition-4
>>> under /0/sa04?
>>
>> You have to find first where is the error.
>>
>
> The first error I see is in the error message i.e. I can mount a device
> under many mountpoints
>
> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sa04
> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sb04
> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sc04
>
> never reports "already mounted", there may well be a limit of how many
> mountpoints I can use but that's above my paygrade.
>
> # umount /0/sc04
>
> leaves only the first 2 mountpoints mounted under
>
> # umount PARTUUID=04edd347-8460-c449-a350-4716a10d0386
>
> unmounts it from the next to last remaining mountpoint
>
> and so on until there are no more.
>
>
> So in response to
>
> mount /dev/sda4 /0/sa04
>
> the message "already mounted OR mount point busy" can only mean that the
> mountpoint (/0/sa04) is busy i.e. some device is already mounted under
> /0/sa04. But /0/sa04 is *my* mountpoint, the system doesn't know it
> exists so

Of course it knows it exists.

>
> 1
> how can it mount anything under it?
>
> 2
> how does it get the authority to mount ANY partition without being
> 'commanded' by root to mount it either in fstab or explicitly?

There are several services and applications that can effect mounts.

>
> lsof doesn't tell me anything significant
>
> and your lsblok one-liner (thanks BTW) just ends with
>
> "[14551.227219] /dev/sda4: Can't open blockdev"

Is it really a blockdev? Do check it.

ls -l /dev/sda*

>
> Maybe THAT's the problem, "the criminal is using dev-names!!!!"
>
> Well, I'nm NOT going to start copy-pasting two-yard long UUIDS just
> becasue the NSA wanted absolute traceability on everything including La
> Plume de Ma Tante.

LOL! The NSA? Really?

> As an extreme worst case scenario (should I ever end
> up with my back to the wall) I might either label partitions with the
> last VERY-FEW digits of the disk RealHardware model number and partition
> number, or I might just boot Slackware and do 'mount /dev/sda4 /0/sa04'.
>
>

--
Cheers, Carlos.

On 3/31/23 20:17, Carlos E.R. wrote:
> On 2023-04-01 01:46, bad sector wrote:
>> On 3/31/23 16:27, Carlos E.R. wrote:
>>> On 2023-03-31 21:03, bad sector wrote:
>>>> On 3/31/23 07:41, Carlos E.R. wrote:
>>>>> lsblk --output
>>>>> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S
>>>>
>>>> Cool, but it doesn't answer two of my questions :-)
>>>>
>>>> 1
>>>> In the attempted command "mount  /dev/sda4 /0/sa04" the
>>>> 'mountpoint' was /0/sa04. The error message stating that
>>>> THE mountpoint was in use is BS needing to be both clarified
>>>> AND fixed.
>>>
>>> I pose that pretense is just an error in the blob that sits between
>>> the chair and the computer :-P  :-D
>>>
>>> If the message says the mount point is in use, then it is. Actually,
>>> the message said two things with and 'or' in between.
>>>
>>>
>>> You can use the command "mount" to see what is actually mounted and
>>> where. And "lsof" to see what is busy.
>>>
>>>
>>>>
>>>> 2
>>>> How do I mount the prinmary (or only) disk partition-4
>>>> under /0/sa04?
>>>
>>> You have to find first where is the error.
>>>
>>
>> The first error I see is in the error message i.e. I can mount a
>> device under many mountpoints
>>
>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sa04
>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sb04
>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sc04
>>
>> never reports "already mounted", there may well be a limit of how many
>> mountpoints I can use but that's above my paygrade.
>>
>> # umount /0/sc04
>>
>> leaves only the first 2 mountpoints mounted under
>>
>> # umount PARTUUID=04edd347-8460-c449-a350-4716a10d0386
>>
>> unmounts it from the next to last remaining mountpoint
>>
>> and so on until there are no more.
>>
>>
>> So in response to
>>
>> mount /dev/sda4 /0/sa04
>>
>> the message "already mounted OR mount point busy" can only mean that
>> the mountpoint (/0/sa04) is busy i.e. some device is already mounted
>> under /0/sa04. But /0/sa04 is *my* mountpoint, the system doesn't know
>> it exists so
>
> Of course it knows it exists.
>
>>
>> 1
>> how can it mount anything under it?
>>
>> 2
>> how does it get the authority to mount ANY partition without being
>> 'commanded' by root to mount it either in fstab or explicitly?
>
> There are several services and applications that can effect mounts.
>
>>
>> lsof doesn't tell me anything significant
>>
>> and your lsblok one-liner (thanks BTW) just ends with
>>
>> "[14551.227219] /dev/sda4: Can't open blockdev"
>
> Is it really a blockdev? Do check it.
>
> ls -l /dev/sda*

# ls -l /dev/sda
brw-rw---- 1 root disk 8, 0 Mar 31 14:39 /dev/sda

# ls -l /dev/sda1
brw-rw---- 1 root disk 8, 1 Mar 31 14:39 /dev/sda1

fdisk ALSO lists all of them in addition to the 'bm' numbers

>> Maybe THAT's the problem, "the criminal is using dev-names!!!!"
>>
>> Well, I'nm NOT going to start copy-pasting two-yard long UUIDS just
>> becasue the NSA wanted absolute traceability on everything including
>> La Plume de Ma Tante.
>
> LOL! The NSA? Really?

I've read that a 'universally unique id' would fill that bill.
Personally I have no dog in the fight but I will NOT copy-paste two-yard
UUIDS :-)

On 2023-04-01 02:51, bad sector wrote:
> On 3/31/23 20:17, Carlos E.R. wrote:
>> On 2023-04-01 01:46, bad sector wrote:
>>> On 3/31/23 16:27, Carlos E.R. wrote:
>>>> On 2023-03-31 21:03, bad sector wrote:
>>>>> On 3/31/23 07:41, Carlos E.R. wrote:
>>>>>> lsblk --output
>>>>>> NAME,KNAME,RA,RM,RO,PARTFLAGS,SIZE,TYPE,FSTYPE,LABEL,PARTLABEL,PTTYPE,MOUNTPOINT,UUID,PARTUUID,WWN,MODEL,ALIGNMENT | less -S
>>>>>
>>>>> Cool, but it doesn't answer two of my questions :-)
>>>>>
>>>>> 1
>>>>> In the attempted command "mount  /dev/sda4 /0/sa04" the
>>>>> 'mountpoint' was /0/sa04. The error message stating that
>>>>> THE mountpoint was in use is BS needing to be both clarified
>>>>> AND fixed.
>>>>
>>>> I pose that pretense is just an error in the blob that sits between
>>>> the chair and the computer :-P  :-D
>>>>
>>>> If the message says the mount point is in use, then it is. Actually,
>>>> the message said two things with and 'or' in between.
>>>>
>>>>
>>>> You can use the command "mount" to see what is actually mounted and
>>>> where. And "lsof" to see what is busy.
>>>>
>>>>
>>>>>
>>>>> 2
>>>>> How do I mount the prinmary (or only) disk partition-4
>>>>> under /0/sa04?
>>>>
>>>> You have to find first where is the error.
>>>>
>>>
>>> The first error I see is in the error message i.e. I can mount a
>>> device under many mountpoints
>>>
>>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sa04
>>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sb04
>>> # mount PARTUUID=04edd347-8460-c449-a350-4716a10d0386  /0/sc04
>>>
>>> never reports "already mounted", there may well be a limit of how
>>> many mountpoints I can use but that's above my paygrade.
>>>
>>> # umount /0/sc04
>>>
>>> leaves only the first 2 mountpoints mounted under
>>>
>>> # umount PARTUUID=04edd347-8460-c449-a350-4716a10d0386
>>>
>>> unmounts it from the next to last remaining mountpoint
>>>
>>> and so on until there are no more.
>>>
>>>
>>> So in response to
>>>
>>> mount /dev/sda4 /0/sa04
>>>
>>> the message "already mounted OR mount point busy" can only mean that
>>> the mountpoint (/0/sa04) is busy i.e. some device is already mounted
>>> under /0/sa04. But /0/sa04 is *my* mountpoint, the system doesn't
>>> know it exists so
>>
>> Of course it knows it exists.
>>
>>>
>>> 1
>>> how can it mount anything under it?
>>>
>>> 2
>>> how does it get the authority to mount ANY partition without being
>>> 'commanded' by root to mount it either in fstab or explicitly?
>>
>> There are several services and applications that can effect mounts.
>>
>>>
>>> lsof doesn't tell me anything significant
>>>
>>> and your lsblok one-liner (thanks BTW) just ends with
>>>
>>> "[14551.227219] /dev/sda4: Can't open blockdev"
>>
>> Is it really a blockdev? Do check it.
>>
>> ls -l /dev/sda*
>
> # ls -l /dev/sda
> brw-rw---- 1 root disk 8, 0 Mar 31 14:39 /dev/sda
>
> # ls -l /dev/sda1
> brw-rw---- 1 root disk 8, 1 Mar 31 14:39 /dev/sda1
>

So, you do not have /dev/sda4

notice I asked for "ls -l /dev/sda*" >:-)

> fdisk ALSO lists all of them in addition to the 'bm' numbers
>
>>> Maybe THAT's the problem, "the criminal is using dev-names!!!!"
>>>
>>> Well, I'nm NOT going to start copy-pasting two-yard long UUIDS just
>>> becasue the NSA wanted absolute traceability on everything including
>>> La Plume de Ma Tante.
>>
>> LOL! The NSA? Really?
>
> I've read that a 'universally unique id' would fill that bill.
> Personally I have no dog in the fight but I will NOT copy-paste two-yard
> UUIDS :-)

Your choice. But if you insist on using names like /dev/sdXY you get the
trouble.

Ah, you can change the UUID to something of your choice...

--
Cheers, Carlos.

On 4/1/23 08:52, Carlos E.R. wrote:
> On 2023-04-01 02:51, bad sector wrote:

>>>>
>>>> and your lsblok one-liner (thanks BTW) just ends with
>>>>
>>>> "[14551.227219] /dev/sda4: Can't open blockdev"
>>>
>>> Is it really a blockdev? Do check it.
>>>
>>> ls -l /dev/sda*
>>
>> # ls -l /dev/sda
>> brw-rw---- 1 root disk 8, 0 Mar 31 14:39 /dev/sda
>>
>> # ls -l /dev/sda1
>> brw-rw---- 1 root disk 8, 1 Mar 31 14:39 /dev/sda1
>>
>
> So, you do not have /dev/sda4
>
> notice I asked for "ls -l /dev/sda*"  >:-)

my mistake

# ls -l /dev/sda*
brw-rw---- 1 root disk 8, 0 Apr 1 12:29 /dev/sda
brw-rw---- 1 root disk 8, 1 Apr 1 12:29 /dev/sda1
brw-rw---- 1 root disk 8, 10 Apr 1 12:29 /dev/sda10
brw-rw---- 1 root disk 8, 11 Apr 1 12:29 /dev/sda11
brw-rw---- 1 root disk 8, 12 Apr 1 12:29 /dev/sda12
brw-rw---- 1 root disk 8, 13 Apr 1 12:29 /dev/sda13
brw-rw---- 1 root disk 8, 14 Apr 1 12:29 /dev/sda14
brw-rw---- 1 root disk 8, 15 Apr 1 12:29 /dev/sda15
brw-rw---- 1 root disk 259, 0 Apr 1 12:29 /dev/sda16
brw-rw---- 1 root disk 259, 1 Apr 1 12:29 /dev/sda17
brw-rw---- 1 root disk 8, 2 Apr 1 12:29 /dev/sda2
brw-rw---- 1 root disk 8, 3 Apr 1 12:29 /dev/sda3
brw-rw---- 1 root disk 8, 4 Apr 1 12:29 /dev/sda4
brw-rw---- 1 root disk 8, 5 Apr 1 12:29 /dev/sda5
brw-rw---- 1 root disk 8, 6 Apr 1 12:29 /dev/sda6
brw-rw---- 1 root disk 8, 7 Apr 1 12:29 /dev/sda7
brw-rw---- 1 root disk 8, 8 Apr 1 12:29 /dev/sda8
brw-rw---- 1 root disk 8, 9 Apr 1 12:29 /dev/sda9

> Your choice. But if you insist on using names like /dev/sdXY you get the
> trouble.
>
> Ah, you can change the UUID to something of your choice...

like so?

tune2fs -U "dev_sda4" /dev/sda4 :-)

On 2023-04-01 18:41, bad sector wrote:
> On 4/1/23 08:52, Carlos E.R. wrote:
>> On 2023-04-01 02:51, bad sector wrote:
>
>>>>>
>>>>> and your lsblok one-liner (thanks BTW) just ends with
>>>>>
>>>>> "[14551.227219] /dev/sda4: Can't open blockdev"
>>>>
>>>> Is it really a blockdev? Do check it.
>>>>
>>>> ls -l /dev/sda*
>>>
>>> # ls -l /dev/sda
>>> brw-rw---- 1 root disk 8, 0 Mar 31 14:39 /dev/sda
>>>
>>> # ls -l /dev/sda1
>>> brw-rw---- 1 root disk 8, 1 Mar 31 14:39 /dev/sda1
>>>
>>
>> So, you do not have /dev/sda4
>>
>> notice I asked for "ls -l /dev/sda*"  >:-)
>
> my mistake
>
> # ls -l /dev/sda*
> brw-rw---- 1 root disk   8,  0 Apr  1 12:29 /dev/sda
> brw-rw---- 1 root disk   8,  1 Apr  1 12:29 /dev/sda1
> brw-rw---- 1 root disk   8, 10 Apr  1 12:29 /dev/sda10
> brw-rw---- 1 root disk   8, 11 Apr  1 12:29 /dev/sda11
> brw-rw---- 1 root disk   8, 12 Apr  1 12:29 /dev/sda12
> brw-rw---- 1 root disk   8, 13 Apr  1 12:29 /dev/sda13
> brw-rw---- 1 root disk   8, 14 Apr  1 12:29 /dev/sda14
> brw-rw---- 1 root disk   8, 15 Apr  1 12:29 /dev/sda15
> brw-rw---- 1 root disk 259,  0 Apr  1 12:29 /dev/sda16
> brw-rw---- 1 root disk 259,  1 Apr  1 12:29 /dev/sda17
> brw-rw---- 1 root disk   8,  2 Apr  1 12:29 /dev/sda2
> brw-rw---- 1 root disk   8,  3 Apr  1 12:29 /dev/sda3
> brw-rw---- 1 root disk   8,  4 Apr  1 12:29 /dev/sda4
> brw-rw---- 1 root disk   8,  5 Apr  1 12:29 /dev/sda5
> brw-rw---- 1 root disk   8,  6 Apr  1 12:29 /dev/sda6
> brw-rw---- 1 root disk   8,  7 Apr  1 12:29 /dev/sda7
> brw-rw---- 1 root disk   8,  8 Apr  1 12:29 /dev/sda8
> brw-rw---- 1 root disk   8,  9 Apr  1 12:29 /dev/sda9

Ok, all seems fine.

Just the other day I could not write to an USB stick because its device
had been replaced with a file, same name. Weird trouble with absurd
error messages similar to yours.

So the system was right in those messages to me. "It" was right, not me.

>> Your choice. But if you insist on using names like /dev/sdXY you get
>> the trouble.
>>
>> Ah, you can change the UUID to something of your choice...
>
> like so?
>
> tune2fs -U "dev_sda4" /dev/sda4   :-)

If you put that on a file or script, then yes, that is asking for trouble.

--
Cheers, Carlos.