https://bugzilla.kernel.org/show_bug.cgi?id=43292

--- Comment #19 from Florian Pritz <flo@xssn.at>  2012-06-06 21:56:31 ---

(In reply to comment #18)

OK, thanks for testing. So you say the problem resolves itself even when you
don't run dd, just it takes longer? I'm more and more inclined to blame some
hardware strangeness... Also originally you said the problem started happening
around 3.3 or 3.4 but now you are testing with 3.0. So do you have any kernel
where you don't see the problem?

So far I can *not* reproduce the problem with 3.0.17 (transfered 8gb of data).
However, with 3.0.32 I can reproduce every time by copying a 20mb+ file.

If you want, I can test a few kernels in between .17 and .32, but that's going
to take some time.

FYI: I do notice short hickups on the client, but those are most likely just
because the disks aren't fast enough to handle the write requests. I'm running
with Gbit ethernet and reading is obviously faster than writing and writing is
slower than the link limit, resulting in the network buffers filling up. I
don't consider this a bug.

-- 
Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching the assignee of the bug.