Re: [gentoo-user] Re: problem formatting new 256 GB USB stick

[Michael <confabulate@kintzios.com>]

[-- Attachment #1: Type: text/plain, Size: 6937 bytes --]

On Sunday 16 February 2025 09:08:10 Greenwich Mean Time Nuno Silva wrote:
> On 2025-02-16, Philip Webb wrote:
> > 250215 Michael wrote:

> >> Formatting a 256GB USB drive, especially if it is a USB 3.0 or later
> >> spec, should not take hours, but minutes if not seconds.
> > 
> > See listing below.  My notes tell me that in many previous cases,
> > it has taken these rates to format :  2 :  6 min/GB  ; 3 :  1,8 min/GB ;
> > today, it took  2 h 51 m  to format a  64 GB  partition (mainly inodes).

I just formatted a USB 2.0 8GB stick with mke2fs as a test.  It took 46 
seconds.  Extrapolating for your 64GB partition it should take ~ 6 minutes.  A 
full 256GB drive would take 24.5 minutes.  I expect a drive enjoying USB 3.0 
transfer speeds to take way less than this.


> > root:635 ~> dmesg -W
> > [2023143.202399] usb 10-2: USB disconnect, device number 2
> > [2023143.210193] blk_print_req_error: 716 callbacks suppressed
> > [2023143.210196] device offline error, dev sdb, sector 95422464 op
> > 0x1:(WRITE) flags 0x800 phys_seg 1 prio class 2 [2023143.210202]
> > buffer_io_error: 1328 callbacks suppressed
> > [2023143.210203] Buffer I/O error on dev sdb1, logical block 11927552,
> > lost async page write [2023143.210218] device offline error, dev sdb,
> > sector 95422472 op 0x1:(WRITE) flags 0x100000 phys_seg 1 prio class 2
> [...]
> 
> > [2023143.210401] device offline error, dev sdb, sector 97257472 op
> > 0x1:(WRITE) flags 0x800 phys_seg 2 prio class 2 [2023143.926864] usb
> > 10-2: new SuperSpeed USB device number 3 using xhci_hcd [2023143.946392]
> > usb 10-2: New USB device found, idVendor=0951, idProduct=1666, bcdDevice=
> > 1.10 [2023143.946397] usb 10-2: New USB device strings: Mfr=1, Product=2,
> > SerialNumber=3 [2023143.946400] usb 10-2: Product: DataTraveler 3.0
> > [2023143.946402] usb 10-2: Manufacturer: Kingston
> > [2023143.946404] usb 10-2: SerialNumber: E0D55EA57410E8B189D80112
> > [2023143.946819] usb-storage 10-2:1.0: USB Mass Storage device detected
> > [2023143.946993] scsi host12: usb-storage 10-2:1.0
> > [2023144.950983] scsi 12:0:0:0: Direct-Access     Kingston DataTraveler
> > 3.0 PMAP PQ: 0 ANSI: 6 [2023144.951249] sd 12:0:0:0: Attached scsi
> > generic sg1 type 0
> > [2023144.951349] sd 12:0:0:0: [sdb] 484540416 512-byte logical blocks:
> > (248 GB/231 GiB) [2023144.951913] sd 12:0:0:0: [sdb] Write Protect is off
> > [2023144.951917] sd 12:0:0:0: [sdb] Mode Sense: 45 00 00 00
> > [2023144.952478] sd 12:0:0:0: [sdb] Write cache: disabled, read cache:
> > enabled, doesn't support DPO or FUA [2023145.369087]  sdb: sdb1 sdb2 sdb3
> > sdb4 < >
> > [2023145.369453] sd 12:0:0:0: [sdb] Attached SCSI removable disk
> > 
> > I can't make sense of the "USB disconnect" at the beginning

This is the message you get when you physically unplug a USB drive.


> > nor the "Attached SCSI removable disk" at the end :

This is the message you get when the device is powered up, detected by the 
kernel and the filesystem is then being accessed.  From this point on the 
device can be read from and written to.


> > I didn't touch anything during the whole process & have no removable
> > disks.

Needless to say, if you did not touch the USB stick when the disconnection 
message was recorded something disconnected (electrically) the device.


> >> If both of these sticks are behaving the same way,
> >> it could be the port on your PC which has a problem.
> >> You can try using a different USB port
> >> to eliminate this causing the formatting failure.
> > 
> > I used the port normally used by my scanner with no problems.
> > The behaviour is the same.

Hmm ... this points to the USB sticks, rather than the ports on your device.


> The device disconnected and then reconnected, this hopefully means there
> is nothing wrong with the flash stick itself, that you just have to
> figure out what is disconnecting it: lack of power, bad cable (if
> applicable), usb power saving...

Wouldn't the same problem manifest with other USB sticks?  That said, I have 
an old USB stick which fails to connect properly if I push it in all the way 
in a port.  I have to push it in, then imperceptibly try to pull it out to 
make good electrical contact with the port.  I've blamed corrosion on the 
electrical strips inside the USB connector - it had been dropped in a mug of 
coffee at some point!  Philip's USB sticks are brand new.


> I've seen USB3 devices be very unstable when connected directly to some
> USB ports. Adding a (USB extension, all of this with USB A) cable solved
> the issue on the machine I saw that on. But that's probably not the
> case, unless what you're seeing is mostly reproducible but random in how
> long it takes to manifest (i.e. time between mkfs and the first
> disconnect message in the kernel log).

I've had a similar problem with a new USB 3.0 stick.  I can't recall the make/
model, but it was not some cheap knock-off.  The speeds were glacial, rsync 
would fail sooner or later when copying data over, disconnections were shown 
on dmesg and the stick would become really hot to touch.  I tried reformatting 
it with FAT, ext2 and finally with F2FS, with improving performance each time 
due to the lighter load placed by the filesystem type on the device.  However, 
the disconnections and failures continued until I returned it as a faulty 
product.  Unlike the more expensive USB SSD drives, the onboard controller of 
the lower cost USB flash sticks tend to be cheap and unsophisticated, with a 
higher probability of early failure.


> >> Other than a hardware problem with the device itself,
> >> there is the chance of counterfeit USB drives, churned out at volume
> >> and having a smaller size and speed than advertised
> >> or such poor quality flash chips they end up corrupting data.
> >> Usually they survive a reformat, at least with FAT,
> >> but can fail at any point thereafter.
> > 
> > That's very unlikely for sticks bought from a reputable store.
> > I've used Canada Computers since 2000 & had a problem only once,
> > a defective mobo for a new machine, which was replaced by another make.

"Reputable store" these days implies they will not try to rob you, they have 
some rudimentary technical knowledge about their products and you can return 
faulty products for a replacement/refund.  Other than that they are all box-
shifters of products typically manufactured in the Far East at minimum cost 
and invariably with very little quality control.

Regarding the USB stick marketed as Kingston Exodia 256 GB, the dmesg shown 
brand and model "idVendor=0951, idProduct=1666" helps identify it as being the 
same with the older Kingston DataTraveler 100 G3/G4/SE9 G2/50:

https://devicehunt.com/view/type/usb/vendor/0951/device/1666

If you intend to return it to the store, first reformat it with FAT and do not 
mention your ext2 experience, in case they try to blame that as the cause of 
your disconnections.

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 833 bytes --]