Re: usb hard drive changes from /dev/sda to /dev/sdb with sarge
On Sat, Dec 10, 2005 at 03:53:34PM -0800, Jason Dunsmore wrote:
> once in a while (can be every couple days or every couple weeks), my
> usb hard drive changes from /dev/sda to /dev/sdb, and i see this in
> dmesg:
>
>
> scsi11 (0:0): rejecting I/O to dead device
> ReiserFS: sda1: warning: zam-7001: io error in reiserfs_find_entry
> scsi11 (0:0): rejecting I/O to dead device
> scsi11 (0:0): rejecting I/O to dead device
> ReiserFS: sda1: warning: vs-13050: reiserfs_update_sd: i/o failure
> occurred tryi
> ng to update [1 2 0x0 SD] stat data
> scsi11 (0:0): rejecting I/O to dead device
> Buffer I/O error on device sda1, logical block 8210
> lost page write due to I/O error on sda1
> ReiserFS: sda1: warning: journal-837: IO error during journal replay
> REISERFS: abort (device sda1): Write error while updating journal
> header in flus
> h_journal_list
> REISERFS: Aborting journal for filesystem on sda1
> scsi11 (0:0): rejecting I/O to dead device
> Buffer I/O error on device sda1, logical block 2480
> lost page write due to I/O error on sda1
> Buffer I/O error on device sda1, logical block 2481
> lost page write due to I/O error on sda1
> scsi11 (0:0): rejecting I/O to dead device
> Buffer I/O error on device sda1, logical block 2482
> lost page write due to I/O error on sda1
> scsi11 (0:0): rejecting I/O to dead device
> scsi11 (0:0): rejecting I/O to dead device
> ReiserFS: sda1: warning: zam-7001: io error in reiserfs_find_entry
> scsi11 (0:0): rejecting I/O to dead device
> ReiserFS: sda1: warning: zam-7001: io error in reiserfs_find_entry
> scsi11 (0:0): rejecting I/O to dead device
> ReiserFS: sdb1: found reiserfs format "3.6" with standard journal
> ReiserFS: sdb1: using ordered data mode
> ReiserFS: sdb1: journal params: device sdb1, size 8192, journal first
> block 18, max trans len 1024, max batch 900, max commit age 30, max
> trans age 30
> ReiserFS: sdb1: checking transaction log (sdb1)
> ReiserFS: sdb1: replayed 3 transactions in 0 seconds
> ReiserFS: sdb1: Using r5 hash to sort names
>
>
> after that happens, i have to remount my usb hard drive under /dev/sdb.
> after a reboot, it's at /dev/sda again. this causes all kinds of
> problems, because i back up to that drive, so if it changes and i don't
> catch it, my root drive gets maxed out. i'm using the prebuilt
> 2.6.8-2-k7 kernel with sarge.
>
> any ideas how i can fix this?
>
No ideas as to how you can fix it, but it appears to be similar symptoms
to what I observe with a digital camera on usb. If I power down the
camera while it is connected via usb, bad things happen with device
assignment, and the device ID that was being used for camera connection
is no longer available. If, on the other hand, I pull the usb cable
while connected and fully powered, all is OK. That is the device is
unmounted cleanly and the device ID can be reused to mount another,
or the same, device.
Could there be a power glitch on your HD? If so and it behaves like
my camera, it could be killing something in the Linux USB implementation.
I have not yet filed a bug on my problem because I don't understand
the kernel well enough to know what part to blame.
In general the device ID assignment logic in USB/SCSI is inexplicable.
It is a puzzle to me how the observed behavior could be the result of
execution of a computer algorithm.
--
Paul E Condon
pecondon@mesanetworks.net
Reply to: