Thanks Tim. I also tried the xfsdump-1.0.11 tarball with the same
results. I'm sending you the output directly. Let me know if you need
> Hi Jim,
> Having a quick look at this code I don't know what
> is happening here. I haven't seen this before.
> However, I'll post an SGI bug and look at the
> code in more detail.
> During the restore you could turn on full debugging
> with -v5 and send me the last 100 lines or so of output
> leading up to the core dump and perhaps the first 100 lines.
> On Tue, Jul 24, 2001 at 10:35:06AM -0400, Jim Eshleman wrote:
> > Greets! I'm receiving the following error using xfsrestore:
> > # xfsrestore -t -f /dev/st0
> > xfsrestore: using online session inventory
> > xfsrestore: searching media for directory dump
> > xfsrestore: reading directories
> > xfsrestore: directory post-processing
> > xfsrestore: reading non-directory files
> > [...long list of files ...]
> > xfsrestore: examining media file 1
> > xfsrestore: seeking past media file directory dump
> > xfsrestore: drive_scsitape.c:1461: do_next_mark: Assertion
> > `rechdrp->first_mark_offset - rechdrp->file_offset <= ( off64_t )
> > ( contextp->dc_recsz )' failed.
> > Aborted (core dumped)
> > #
> > This is happening on two different systems. System A is the stock SGI
> > RH7.1 XFS 1.0.1 distribution running on a dual P2-333 with a DLT4000
> > drive. System B is RH7.1 + kernel 2.4.7 +
> > patch-2.4.7-xfs-2001-07-22.bz2
> > running on a Netfinity x370 with a DLT7000.
> > I've tried specifying a block size of 240K on the dump and restore.
> > Same
> > result.
> > The dump runs with no errors as "xfsdump -l 0 -o -f /dev/st0 -s userdir
> > /a"
> > where /a/userdir is about 1.2G.