xfs
[Top] [All Lists]

Re: xfsrestore problems

To: Charles Shannon Hendrix <shannon@xxxxxxxxxxxxx>
Subject: Re: xfsrestore problems
From: John Kihonge <mkulima@xxxxxxx>
Date: Mon, 20 May 2002 16:20:42 -0500
Cc: linux-xfs@xxxxxxxxxxx
References: <20020511205125.GD4462@widomaker.com> <20020516020938.GB10343@widomaker.com> <3CE3D3CF.7DBAF8A9@sgi.com> <200205162151.07376.hasch@t-online.de> <20020517000623.GA13675@widomaker.com> <3CE53102.2B550CF8@sgi.com> <20020517181451.GC13880@widomaker.com> <3CE91559.D5D36B75@sgi.com> <20020520203135.GA4906@widomaker.com>
Sender: owner-linux-xfs@xxxxxxxxxxx
Charles Shannon Hendrix wrote:

> On Mon, May 20, 2002 at 10:25:14AM -0500, John Kihonge wrote:
>
> > > xfsrestore: examining media file 3
> > > xfsrestore: seeking past media file directory dump
> > > xfsrestore: drive_scsitape.c:1507: do_next_mark: Assertion
> > > `rechdrp->first_mark_offset - rechdrp->file_offset <= ( off64_t )(
> > > contextp->dc_recsz )' failed.
> > > Aborted
> > >
> >
> > If the problem is just the assertion one, it maybe that you are restoring an
> > old dump (dumped prior to a fix for xfsdump-1.1.10 of 10th December 2001).
> >
> > The mail archive:
> >   http://marc.theaimsgroup.com/?l=linux-xfs&m=101435725816823&w=2 has an
> >   explanation of the problem and the suggested fix to allow one to restore
> >   such a dump.
> >
> > I hope that will fix the problem.
>
> I made the recommended change to common/xlate.c, removing the
> first_mark_offset endian conversion.
>
> No joy:
>
> xfsrestore: restoring non-directory files
> xfsrestore: examining media file 8
> xfsrestore: seeking past media file directory dump
> xfsrestore: drive_scsitape.c:1481: do_next_mark: Assertion
> `rechdrp->first_mark_offset != 0' failed.
>
> I then made the recommended changes to 2.01 of xfsdump drive_scsitape.c.
> The change was for 1.5, but the file looks about the same.  The
> modification goes just before the failing assertion listed in the
> error above.

Try making the recommended changes to

    restore/arch_xlate.c

and see if it makes any difference. If it dosnt work, use -v5 option and put the
output in a location that we could look at it to see if we can find any
indication of what is hapenning.

>
>
> The results from running this version is:
>
> ...the same exact error.
>
> --
> UNIX/Perl/C/Pizza__________________________________shannon@xxxxxxxxxxxxx

----
Kihonge JN


[[HTML alternate version deleted]]


<Prev in Thread] Current Thread [Next in Thread>