Re: [Cooker] Copying of large files to flashdisk fails in Cooker

[Steve Morris]

Frank Griffin wrote:
> Steve Morris wrote:
>> I am now totally confused. I tried the process again just now under kde4
>> with krusader  after issuing the command 'mount -tauto /dev/sda1
>> /mnt/flashdisk' as root and the copy process worked fine. This now says
>> the failure is random, so what the h**l is going on. When I get the
>> issue when using krusader, if I click on the 'cancel' button on the
>> krusader copy dialogue, it terminates the copy process but has no impact
>> on what the flashdisk is doing. The only way I have been able interrupt
>> the flashdisk process is to physically unplug the flashdisk, which so
>> far has not corrupted the contents. I cannot be certain but it also
>> appears that when I get this error and disconnect the flashdisk I am
>> unable to shut the system down, it drops back to the blue background
>> screen and just sits there in busy state.
>>   
> It would really simplify matters if you would
>
> 1) eliminate the GUI and use cp (or even better, dd, since you can use
> kill -SIGUSR1 to see where it's at)
>
> 2) dig up the log messages associated with the stall.
>
> The behavior you describe could result from a lot of things.  If
> krusader uses a really large buffer for the copy, you could have
> front-loaded the kernel driver with enough data in the early part of
> the copy to keep the device busy for quite a while whether krusader
> quits or not.  An advantage of dd is that you can specify the
> blocksize.  Overwhelming the driver could have triggered a driver bug
> which manifests as a loop, or perhaps the driver is taking data too
> fast and depleting kernel memory to hold it all.  I don't know, and
> you don't know, but syslog probably does :-)
>
> The test I ran was from a terminal window under GNOME, using a Cruzer
> Micro 2.0GB SanDisk.  Assuming we're using the same kernel driver and
> have similar amounts of RAM (2GB here), my copy ran pretty fast.
>
> My syslog contents for the duration were:
>
> May  1 10:14:32 ftgme2 kernel: usb 1-5: new high speed USB device
> using ehci_hcd and address 4
> May  1 10:14:32 ftgme2 kernel: usb 1-5: configuration #1 chosen from 1
> choice
> May  1 10:14:33 ftgme2 chipcardd[4242]: devicemanager.c: 3356: Changes
> in hardware list
> May  1 10:14:33 ftgme2 kernel: Initializing USB Mass Storage driver...
> May  1 10:14:33 ftgme2 kernel: scsi2 : SCSI emulation for USB Mass
> Storage devices
> May  1 10:14:33 ftgme2 kernel: usbcore: registered new interface
> driver usb-storage
> May  1 10:14:33 ftgme2 kernel: USB Mass Storage support registered.
> May  1 10:14:33 ftgme2 kernel: usb-storage: device found at 4
> May  1 10:14:33 ftgme2 kernel: usb-storage: waiting for device to
> settle before scanning
> May  1 10:14:38 ftgme2 kernel: scsi 2:0:0:0: Direct-Access    
> SanDisk  U3 Cruzer Micro  3.27 PQ: 0 ANSI: 2
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] 4001422 512-byte
> hardware sectors (2049 MB)
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Write Protect is off
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Mode Sense: 03 00 00 00
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Assuming drive cache:
> write through
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] 4001422 512-byte
> hardware sectors (2049 MB)
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Write Protect is off
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Mode Sense: 03 00 00 00
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Assuming drive cache:
> write through
> May  1 10:14:38 ftgme2 kernel:  sdd: sdd1
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: [sdd] Attached SCSI
> removable disk
> May  1 10:14:38 ftgme2 kernel: sd 2:0:0:0: Attached scsi generic sg3
> type 0
> May  1 10:14:38 ftgme2 kernel: scsi 2:0:0:1: CD-ROM           
> SanDisk  U3 Cruzer Micro  3.27 PQ: 0 ANSI: 2
> May  1 10:14:38 ftgme2 kernel: scsi 2:0:0:1: Attached scsi generic sg4
> type 5
> May  1 10:14:38 ftgme2 kernel: usb-storage: device scan complete
> May  1 10:14:39 ftgme2 kernel: Driver 'sr' needs updating - please use
> bus_type methods
> May  1 10:14:39 ftgme2 kernel: sr0: scsi3-mmc drive: 8x/40x writer
> xa/form2 cdda tray
> May  1 10:14:39 ftgme2 kernel: sr 2:0:0:1: Attached scsi CD-ROM sr0
> May  1 10:14:40 ftgme2 hald: mounted /dev/sdd1 on behalf of uid 501
> May  1 10:21:24 ftgme2 hald: unmounted /dev/sdd1 from '/media/disk' on
> behalf of uid 501
>
Frank I have 1GB ram and am using an AData Show Me Disk 2.0GB (This has
a graphic capacity display window on the disk). I tried DD last night
and let it continue trying to write to the device for an hour before
trying to kill DD, which GPS was not able to do via either of its kill
methods (hangup or sig). The only way to stop the process was to force
close the shell window DD was launched in and physically disconnect the
flashdisk, this also prevented the machine from being able to perform a
clean shutdown as the only way I was able to shutdown was to use the
power off button.
I checked /var/log/syslog while to operation was in progress and the
only message I get different related to the flashdisk that are different
to yours are:

    My flashdisk is being mounted as sda sda1 as opposed to yours of sdd
    sdd1.
    I get the message Driver 'sd' needs updating - please use bus_type
    methods
    I don't remember seeing the hald: mounted /dev/sda1 message, as
    being under kde4 I have to manually mount the device using 'mount
    -tauto /dev/sda1 /mnt/flashdisk' under root as kde4 doesn't hotplug
    in Mandriva (I haven't determined yet whether it is a Mandriva issue
    or kde), whereas kde3 hotplugs fine, but I also get the issue under
    kde3 as well.

regards,
Steve

begin:vcard
fn:Steve Morris
n:Morris;Steve
email;internet:samorris@netspace.net.au
x-mozilla-html:TRUE
version:2.1
end:vcard