[Bugs-team] kernel panic when mounting dedup dataset

[Brandon Trussell]

Hi All,

I have been experiencing kernel panics 100% of the time when mounting
one dataset in particular.  I want to file a bug on this, but as I'm
new to bug reporting on this project, I figured I'd get some better
instructions from this mailing list.

History / What happened
===================
The box would crash during writing a 4GB file to the dedup dataset via NFS.
Each time the box booted normally afterward, and seemed to work
normally, until I would attempt the copy again.
I attempted to turn off dedup on that dataset, but accidentally turned
it off on a different dataset.
I attempted the copy again, it crashed, but this time it wouldn't boot
successfully anymore (kernel panic during every boot attempt).
Today I determined the entire pool is not damaged, only that dataset
that I was copying to, therefore I have access to the full OS again
and the crash dump.

I have been experiencing the exact same crash ever since I made that
dedup change (same stack trace and line of code outlined by the panic
output).
The crash dump is about 440MB.
Please let me know where I should upload it and anything else I should
include in the bug report.

Here is a look at the panic output.
http://brandon.io/photo-1.JPG

P.S. I'm interesting in trying to role back the transactions on this
zpool... hopefully to get access to my dataset which refuses to mount.
 If there is a process available to get this dataset working again,
please let me know.

Thanks,
Brandon