Strange issue with Samba on 8.0rc1
Pyun YongHyeon
pyunyh at gmail.com
Fri Oct 23 18:37:11 UTC 2009
On Thu, Oct 22, 2009 at 03:29:51PM -0700, David Ehrmann wrote:
> Ivan Voras wrote:
> >This looks like too much problems at the same time to be a problem in
> >samba or FreeBSD.
> >
> >For what it's worth, here are some ideas and data points:
> >
> >- FreeBSD 8.0-RC1 installs and runs perfectly normal in VMWare ESXi
> >with LSI emulation
> >- If it is a disk problem (bad cable? controller?) it could manifest
> >itself or at least give some clue if you did 'zpool scrub'. ZFS is
> >checksummed - it should give you IO errors if the data gets corrupted.
> >- There are several versions of Samba in the ports - can you try
> >another one?
> >- Before you do anything, have you updated to the latest 8.0 source?
>
> Odd. Trying to rebuild world, csup failed with "Receiver: Connection
> reset by peer." cvsup didn't do any better. I didn't have this problem
> with the VM (where I'm actually going to build everything, then copy it
> over). Could a buggy network driver cause this problem? It's a Via
> gigabit ethernet (vge) device.
>
I vaguely remember some vge(4) users who have VT613x PCIe variants
reported driver instabilities. So I guess there are some silicon
changes in newer VT613x controllers. If you're using PCIe version
of the controller you may see the same issue. If you have a spare
NIC or USB NIC how about trying other ethernet driver to rule out
possible vge(4) issues?
I managed to find a site in China that sells a stand-alone PCIe
VT6130 card and ordered it but it was not delivered yet. Don't know
how long it would take.
More information about the freebsd-current
mailing list