hi,
On Mon, Sep 02, 2002 at 08:06:25AM -0600, Michael Best wrote:
> Well I initially installed via Gentoo, which I believe was based on 1.1
> code (April 2002).
OK, so I expect that version would fail with the current tools.
(but that is not kernel 2.4.19+20020825 though, right?)
> I was just hoping to dump and restore this filesystem onto an identical
> adjacent (move from hda8 to hda7) in order to grow hda7 to the entire
> size of the two partitions.
>
> But I guess either I didn't compile the utils right or the filesystem on
> disk is still in the older V1 format.
This problem is more to do with ioctl versions that ondisk formats.
> This if for no other reason would be a good reason for a point release
> that contains the V2Log code, if the future utilities aren't going to be
> able to cope with V1. Or is there an easy way to convert?
No other way to convert the ondisk format - mkfs is it; but this is
not an ondisk format problem, its an ioctl version problem. We can
do a better job than we currently are in this kernel <-> userspace
interaction - I will checkin some changes later today which should
allow all the tools to work with both new and old kernels.
cheers.
> Nathan Scott wrote:
> > On Mon, Sep 02, 2002 at 03:33:46AM -0600, Michael Best wrote:
> >
> >>Nathan Scott wrote:
> >>
> >>>Basically, if you compile xfsdump against the xfsprogs headers
> >>>from xfsprogs in 1.1 it should work, but if you compile xfsdump
> >>>against CVS header files it will use the new geometry ioctl,
> >>>which was not supported by the 1.1 kernel code and hence the
> >>>"Inappropriate ioctl" error.
> >>
> >>Odd I had the same error. I am using a 2.4.19+20020825 kernel and the
> >>latest tools. Do I need older tools, an older kernel, or a newer kernel
> >>still? :)
> >
> >
> > Hmm... are you sure you are running that recent a kernel?
> > v2 log support went in on 2002/06/18 and I don't understand
> > how this error would be generated from a current kernel...
> > perhaps I'm overlooking something though. My tests don't
> > seem to produce any errors of this kind.
> >
> > cheers.
>
>
--
Nathan
|