[Novalug] disk wipe causing panic

Mark Smith mark@winksmith.com
Tue Jun 17 17:58:45 EDT 2014


a few comments on my progress....

to whomever pushed me to check for memory, good work.  i'm not actually
believing it, but i am getting literally thousands of memory errors.
so much so that i can't believe that this system was working at all.
that makes me think that there's something wrong with the memory
checker though.  both of the 8GB DIMM's was so tightly pushed in that i
couldn't even reseat them.  oddly, only one side was able to be undone.
i'm afraid to push too hard on the release catch, but one side just
doesn't want to come up.  weird.  i'll need to figure that out so i can
see if it's both DIMM's or just one.

also, the ultimatebootcd (http://www.ultimatebootcd.com/download.html)
is well worth having a CDROM lying around.  it's a collection of lots
of tools to check out your system.  it includes Darik's Boot and Nuke
(dban--http://sourceforge.net/projects/dban/) which is working fine
clearing my drives.  i'm only going for the zero fills, but it offers lots
of other clearing specifications.  if offers different PNRG's as well.
it's clearly a linux based solution even though it doesn't say so.

also, if i hit enter during booting, it gives me lots of choices and
doesn't automatically whack all the disks.

oddly, if i mount the disk, put a file system on it and then create a
file as large as the file system filled with zeroes, it doesn't crash.


On Tue, Jun 17, 2014 at 08:28:21AM +1200, Mark Smith wrote:
> i have a bunch of USB disks i'm trying to wipe.  i have resorted to my
> old friend dd for this.
> 
> 	dd if=/dev/zero of=/dev/sdc bs=1M
> 
> you'd think that'd be the end of this story, but it's just the start.
> my system panics have running for an hour or so.  I have a new-ish
> motherboard with new-ish i5 hardware.  this system is known to work
> fine.  i just took these disks out of a working storage system.  they
> are also known to work.
> 
> i've tried /dev/sdc and /dev/sdc1 from both front/back USB ports with
> the same results.  it either hangs hard or i panic.  when i do get any
> messages, i get a bunch of IO errors from the device i'm writing a stack
> trace with function calls that look like block device writers and a hard
> lock up.
> 
> 	blk_update_request
> 	blk_update_bidi
> 	blk_end_bidi
> 	blk_end_request_err
> 	scsi_io_completion
> 	scsi_finish_command
> 	scsi_softirq_done
> 	blk_done_softirq
> 	__do_softirq
> 	run_ksoftirqd
> 	smpboot_thread_fn
> 	? SyS_setgroups
> 	kthread
> 	? kthread_create_on_node
> 	ret_from_fork
> 	? ktthread_create_on_node
> 	Code: 47 c1 c1 ea 03 83 c2 01 39 d0 0f 47 c2 c3 66 90 0f 1f 44 00 00 85 f6 74 17 f0 80 67 18 fe 48 8b 47 40 48 85 c0 74 07 55 48 89 e5 <ff> d0 5d f3 c3 48 8b 47 18 a8 01 b8 fb ff ff ff 0f 44 f0 eb de
> 
> for the full traceback panic screen:
> 
> 	https://drive.google.com/file/d/0B3jQ8Q3Zo2xcbmhrbDc5LWVyMVE/edit?usp=sharing
> 
> i didn't want to kill my system, so i have it booted off of a USB key.
> not sure what version it is, but it did the same thing when i booted
> normally into my 12.04.4LTS latest stable kernel.
> 
> nobody is into linux around these parts so i'm putting an ntfs filesystem
> on it.  the mkfs.ntfs attempts to do a disk write of every block by
> default.  this also panics my system.
> 
> my last stab effort that i'm trying right now is i run dd for a couple
> gig, re-write the partitions using fdisk, create an ext4 filesystem
> on it, mount, and then create a file which consumes the entire disk.
> finally, i overwrite the ext4 with an ntfs file system in quick mode.
> rinse, repeat.  i expect this to work as i know ext4 works.
> 
> 1. what's going on?
> 2. what's the best way to clear these disks?
> 3. have i really cleared the entire disk?  i was hoping to write to
>    /dev/sdc whacking the entire drive in the process.

-- 
Hei konā mai
Mark Smith
mark@winksmith.com



More information about the Novalug mailing list