[gentoo-user] Why isn't this SDcard mounting?

[Stroller]

I can't make any sense of it.

It's doing this with 2 different SDcards, I think - this is the brand new replacement for one which I believe(d) to be knackered.

Instead of mounting the SDcard, it's mounting the loopback device.

A card, as /dev/sdb, was previously zeroed over and repartitioned a couple of days ago. I bet if I reboot the system it'll be recognised. How do I get it to be recognised now?

Thanks in advance for your thoughts,

Stroller.


$ sudo grep kernel /var/log/messages
Aug 29 13:49:46 alrai kernel: usb 1-1: new high-speed USB device number 5 usingehci-pci
Aug 29 13:49:46 alrai kernel: usb 1-1: New USB device found, idVendor=0bda, idProduct=0136
Aug 29 13:49:46 alrai kernel: usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
Aug 29 13:49:46 alrai kernel: usb 1-1: Product: USB2.0-CRW
Aug 29 13:49:46 alrai kernel: usb 1-1: Manufacturer: Generic
Aug 29 13:49:46 alrai kernel: usb 1-1: SerialNumber: 20060413092100000
Aug 29 13:49:46 alrai kernel: usb-storage 1-1:1.0: USB Mass Storage device detected
Aug 29 13:49:46 alrai kernel: scsi host6: usb-storage 1-1:1.0
Aug 29 13:49:47 alrai kernel: scsi 6:0:0:0: Direct-Access Generic- Card Reader 1.00 PQ: 0 ANSI: 0 CCS
Aug 29 13:49:47 alrai kernel: sd 6:0:0:0: Attached scsi generic sg1 type 0
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] 62521344 512-byte logical blocks: (32.0 GB/29.8 GiB)
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] Write Protect is off
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] Mode Sense: 03 00 00 00
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] No Caching mode page found
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] Assuming drive cache: write through
Aug 29 13:49:48 alrai kernel: sdb: sdb1
Aug 29 13:49:48 alrai kernel: sd 6:0:0:0: [sdb] Attached SCSI removable disk
Aug 29 14:47:12 alrai kernel: sdb: sdb1
$
$ sudo parted /dev/sdb p
Model: Generic- Card Reader (scsi)
Disk /dev/sdb: 32.0GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags:

Number Start End Size Type File system Flags
1 4194kB 32.0GB 32.0GB primary fat32 lba

$ sudo mount -v /dev/sdb1 /mnt/tmp
mount: /dev/loop0 mounted on /mnt/tmp.
$ du -sh /mnt/tmp
10M /mnt/tmp
$ lsmod | grep fat
vfat 20480 1
fat 65536 1 vfat
$

[Ian Zimmerman]

On 2017-08-29 14:53, Stroller wrote:

> $ sudo parted /dev/sdb p
> Model: Generic- Card Reader (scsi)
> Disk /dev/sdb: 32.0GB
> Sector size (logical/physical): 512B/512B
> Partition Table: msdos
> Disk Flags:
>
> Number Start End Size Type File system Flags
> 1 4194kB 32.0GB 32.0GB primary fat32 lba
>
> $ sudo mount -v /dev/sdb1 /mnt/tmp
> mount: /dev/loop0 mounted on /mnt/tmp.

I don't have a quick solution, but I would look at the state of /dev
(not only /dev/sdb* but also the various /dev/disk/by-* directories)
both before and after running parted. parted is my prime suspect for
messing things up here.

Also, is this the normal mount program from util-linux package, or some
"modern" replacement?

--
Please don't Cc: me privately on mailing lists and Usenet,
if you also post the followup to the list or newsgroup.
Do obvious transformation on domain to reply privately _only_ on Usenet.

[Stroller]

> On 29 Aug 2017, at 15:53, Ian Zimmerman <i...@very.loosely.org> wrote:
>
> I don't have a quick solution, but I would look at the state of /dev
> (not only /dev/sdb* but also the various /dev/disk/by-* directories)
> both before and after running parted. parted is my prime suspect for
> messing things up here.

Indeed.

No sdb1 is mentioned, despite it apparently being recognised by the kernel when plugged in (from the last line of the `grep kernel /var/log/messages` output in my previous post).

$ ls -l /dev/disk/* | grep sdb
lrwxrwxrwx 1 root root 9 Aug 29 14:51 usb-Generic-_Card_Reader_20060413092100000-0:0 -> ../../sdb
lrwxrwxrwx 1 root root 9 Aug 29 14:51 pci-0000:00:12.2-usb-0:1:1.0-scsi-0:0:0:0 -> ../../sdb
$

The same command, grepping sda, shows much longer output, with symlinks to all the partitions

Is it udev that's responsible for populating the dev nodes?
(is that the right terminology?)

How do I force it to reconstruct the partition table? Surely one should expect to be able to format or partition a removable drive and have the dev nodes created without the necessity of rebooting?

> Also, is this the normal mount program from util-linux package, or some
> "modern" replacement?

It's the normal one:

$ equery belongs `which mount`
* Searching for /bin/mount ...
sys-apps/util-linux-2.28.2 (/bin/mount)
$

Stroller.

[Alan McKinnon]

On 29/08/2017 17:33, Stroller wrote:
>
>> On 29 Aug 2017, at 15:53, Ian Zimmerman <i...@very.loosely.org> wrote:
>>
>> I don't have a quick solution, but I would look at the state of /dev
>> (not only /dev/sdb* but also the various /dev/disk/by-* directories)
>> both before and after running parted. parted is my prime suspect for
>> messing things up here.
>
> Indeed.
>
> No sdb1 is mentioned, despite it apparently being recognised by the kernel when plugged in (from the last line of the `grep kernel /var/log/messages` output in my previous post).
>
> $ ls -l /dev/disk/* | grep sdb
> lrwxrwxrwx 1 root root 9 Aug 29 14:51 usb-Generic-_Card_Reader_20060413092100000-0:0 -> ../../sdb
> lrwxrwxrwx 1 root root 9 Aug 29 14:51 pci-0000:00:12.2-usb-0:1:1.0-scsi-0:0:0:0 -> ../../sdb
> $
>
> The same command, grepping sda, shows much longer output, with symlinks to all the partitions
>
> Is it udev that's responsible for populating the dev nodes?
> (is that the right terminology?)
>
> How do I force it to reconstruct the partition table? Surely one should expect to be able to format or partition a removable drive and have the dev nodes created without the necessity of rebooting?

run partprobe and see if that makes a difference. It forces the kernel
to re-organize it's idea of what partitions are available.

I would have thought SD Cards were treated like regular hotpluggable
devices like USB storage, but maybe not. I'd be interested to see the
results of running partprobe.


--
Alan McKinnon
alan.m...@gmail.com

[Stroller]

> On 29 Aug 2017, at 16:35, Alan McKinnon <alan.m...@gmail.com> wrote:
>>
>> Is it udev that's responsible for populating the dev nodes?
>> (is that the right terminology?)
>>
>> How do I force it to reconstruct the partition table? Surely one should expect to be able to format or partition a removable drive and have the dev nodes created without the necessity of rebooting?
>
> run partprobe and see if that makes a difference. It forces the kernel
> to re-organize it's idea of what partitions are available.
>
> I would have thought SD Cards were treated like regular hotpluggable
> devices like USB storage, but maybe not. I'd be interested to see the
> results of running partprobe.

$ sudo partprobe -s
/dev/sda: gpt partitions 1 2 3 4 5
/dev/sdb: msdos partitions 1
$

The following is also dumped to /var/log/messages:

Aug 29 17:31:13 alrai sudo[20565]: stroller : TTY=pts/1 ; PWD=/home/stroller ; USER=root ; COMMAND=/bin/bash -c partprobe
Aug 29 17:31:13 alrai sudo[20565]: pam_unix(sudo:session): session opened for user root by (uid=0)
Aug 29 17:31:13 alrai kernel: sdb: sdb1
Aug 29 17:31:13 alrai kernel: sdb: sdb1
Aug 29 17:31:13 alrai sudo[20565]: pam_unix(sudo:session): session closed for user root

However no new device nodes are added in /dev.

This is a headless system, mostly used as a file server. It doesn't run a desktop (although I've run X11 apps using xpra a few times in the past). I've never done anything to set up hotplugging.

Stroller.

[Mick]

This may have been mentioned already, but do you have sys-fs/udisks installed?

Check the output of udisksctl status/monitor/info and see what it reveals.
Then check if you can mount the device with udiskctl.
--
Regards,
Mick

[Stroller]

> On 29 Aug 2017, at 19:15, Mick <michael...@gmail.com> wrote:
> ...

> This may have been mentioned already, but do you have sys-fs/udisks installed?

I did not.

> Check the output of udisksctl status/monitor/info and see what it reveals.
> Then check if you can mount the device with udiskctl.

Having installed it I got "Error connecting to the udisks daemon: Could not connect: No such file or directory" a few times, until it occurred to me to start dbus. This may be the first time I've ever done so.

I was then able to get a status and info for the drive.

I cannot mount the drive using `sudo mount -v /dev/sdb1 /mnt/tmp`, but I am able to do so using `udisksctl`:

$ sudo udisksctl mount -b /dev/sdb1
Mounted /dev/sdb1 at /run/media/root/6D18-12B4.
$

Unexpectedly I find 10MB of files on the device - the same size as the loopback device which previously claimed to be mounted (and 6MB less than the files I had intended to copy to it).

No /dev/sdb1 has been created.

There are aspects of this I still don't understand, but I am grateful to everyone who has provided their time.

Stroller.

[Mick]

On Tuesday, 29 August 2017 22:27:36 BST Stroller wrote:
> > On 29 Aug 2017, at 19:15, Mick <michael...@gmail.com> wrote:
> > ...
> > This may have been mentioned already, but do you have sys-fs/udisks
> > installed?
> I did not.
>
> > Check the output of udisksctl status/monitor/info and see what it reveals.
> > Then check if you can mount the device with udiskctl.
>
> Having installed it I got "Error connecting to the udisks daemon: Could not
> connect: No such file or directory" a few times, until it occurred to me to
> start dbus. This may be the first time I've ever done so.
>
> I was then able to get a status and info for the drive.
>
> I cannot mount the drive using `sudo mount -v /dev/sdb1 /mnt/tmp`, but I am
> able to do so using `udisksctl`:
>
> $ sudo udisksctl mount -b /dev/sdb1
> Mounted /dev/sdb1 at /run/media/root/6D18-12B4.
> $

OK, this shows your kernel modules for mounting this device are not missing
anything critical.

> Unexpectedly I find 10MB of files on the device - the same size as the
> loopback device which previously claimed to be mounted (and 6MB less than
> the files I had intended to copy to it).
>
> No /dev/sdb1 has been created.
>
> There are aspects of this I still don't understand, but I am grateful to
> everyone who has provided their time.
>
> Stroller.

Many desktops use udisks to handle plugging/unplugging:

https://wiki.gentoo.org/wiki/Udisks

which requires dbus and polkit. However, if this is running headless and
minimal without many of the desktop nice-to-have functionality, at least check
you have added your user to the plugdev group and perhaps usb.

Instead of udisks you can also use:

- 'udevadm monitor' and plug your device to see what kernel and udev events it
generates.

- 'lsblk -o +fstype,label,uuid,partuuid' to make sure the device mount point
is not being claimed by some other mount like loopback - which I noticed in
your output, but can't explain why it is happening ... :-/

HTH.
--
Regards,
Mick

[Stroller]

> On 29 Aug 2017, at 14:53, Stroller <stro...@stellar.eclipse.co.uk> wrote:
> ...

> Instead of mounting the SDcard, it's mounting the loopback device.
>
> A card, as /dev/sdb, was previously zeroed over and repartitioned a couple of days ago. I bet if I reboot the system it'll be recognised. How do I get it to be recognised now?

At some point yesterday, making no progress with `udisksctl` (I think that mounted the loopback device, too, despite claiming to have succeeded in mounting /dev/sdb1 as described in my email of 10:20pm 29-8-2017), I just wanted to use the SDcard and rebooted the system.

As expected, when I put the card back in, everything worked perfectly.

My belief is that this is related to zeroing out or recreating the partition table.

Stroller.