That gave me enough of a hint to go on. It does look like it was an
issue with how I originally specified the options.
For posterity's sake, here's how I fixed it:
zfs set sharenfs='-ro -alldirs -network 172.16.1.0 -mask 255.255.255.0'
rpool/data/build/pub
(Applied the same sharenfs settings wi
So... I've been tracking FreeBSD (generally, daily) on various machines
for a couple of decades or so. And I learned long ago to rebuild kmods
from ports whenever I rebuilt the kernel, and that generally, the
simplest way to ensure that that is done is to augment /etc/src.conf
with (e.g.):
PORTS_
On Thu, 24 Oct 2024 at 14:26, Michael Butler
wrote:
>
> It seems there are some additional constraints about non-existent
> directories that now count as errors ..
I ran into this as well, in my work tree that has drm-kmod 6.6 as a
submodule. I think these include paths are just leftover from so
Hey all,
I'm exposing some ZFS datasets using the `zfs set sharenfs` option.
However, this causes mountd to complain:
BEGIN mountd LOG
Oct 24 20:19:59 hbsd-os-build-01 mountd[89840]: bad exports list line
'/data/build/pub/current -ro'
Oct 24 20:20:07 hbsd-os-build-01 mountd[14071]: =
I think that's just an exports(5) syntax error. I don't think "="
should follow "-ro". Try this instead:
zfs set sharenfs='-ro -alldirs 172.16.1.0/24' rpool/data/build/pub
On Thu, Oct 24, 2024 at 2:26 PM Shawn Webb wrote:
>
> Hey all,
>
> I'm exposing some ZFS datasets using the `zfs set share
On October 24, 2024 9:25:35 PM GMT+03:00, Michael Butler
wrote:
>It seems there are some additional constraints about non-existent directories
>that now count as errors ..
>
>--- sync_file.o ---
>cc -O2 -pipe -fno-strict-aliasing -DLINUXKPI_VERSION=60100
>'-DKBUILD_MODNAME="dmabuf"' -DCONFIG_D
