>Number:         25967
>Category:       bin
>Synopsis:       dump -a not continuing after end of media
>Confidential:   no
>Severity:       serious
>Priority:       low
>Responsible:    bin-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jun 19 05:59:00 UTC 2004
>Closed-Date:
>Last-Modified:
>Originator:     Daniel Carosone
>Release:        NetBSD 2.0E
>Organization:
rare
>Environment:
System: NetBSD schizo 2.0E NetBSD 2.0E (_schizo_) #81: Sun May 23 07:57:56 EST 2004 dan@schizo:/home/NetBSD/obj/p3/home/NetBSD/HEAD/src/sys/arch/i386/compile/_schizo_ i386
Architecture: i386
Machine: i386

/sbin/dump:
     $NetBSD: crt0.c,v 1.13 2003/07/26 19:24:27 salo Exp $
     $NetBSD: itime.c,v 1.15 2003/08/07 10:04:14 agc Exp $
     $NetBSD: main.c,v 1.57 2004/04/21 01:05:32 christos Exp $
     $NetBSD: optr.c,v 1.33 2004/04/21 01:05:32 christos Exp $
     $NetBSD: dumprmt.c,v 1.30 2003/08/07 10:04:14 agc Exp $
     $NetBSD: rcache.c,v 1.21 2003/09/13 10:59:50 simonb Exp $
     $NetBSD: tape.c,v 1.43 2004/03/15 01:06:06 lukem Exp $
     $NetBSD: traverse.c,v 1.43 2004/03/24 12:28:51 hannken Exp $
     $NetBSD: unctime.c,v 1.15 2003/08/07 10:04:15 agc Exp $
     $NetBSD: ffs_inode.c,v 1.16 2004/03/27 12:59:18 dsl Exp $
     $NetBSD: ffs_bswap.c,v 1.27 2003/12/31 19:33:13 dbj Exp $

st0 at scsibus0 target 6 lun 0: <Quantum, DLT4000, CD5A> tape removable
st0: density code 26, variable blocks, write-enabled
st0: sync (100.00ns offset 15), 8-bit (10.000MB/s) transfers

>Description:

dumping out a filesystem larger than a single tape:

 # dump 0auhbtf 0 32 /dev/rst0 /media

I get a write error at about the expected point for when the first
tape would be full.

  [14:20:35 EST] DUMP: 41.73% done, finished in 5:14 (at 19:34:48 EST)
  [14:25:35 EST] DUMP: 42.64% done, finished in 5:09 (at 19:35:01 EST)
  [14:28:40 EST] DUMP: write error 18584896 blocks into volume 1

But dump doesn't want to continue onto the next tape:

  DUMP: Do you want to restart?: ("yes" or "no") yes
  [14:53:16 EST] DUMP: Closing this volume.  Prepare to restart with new media;
  [14:53:16 EST] DUMP: this dump volume will be rewritten.

(saying no produces different, but equally non-continuing-on-the-next-tape
noises.)

I assume dump(8) isn't handling this properly, and I'd want it to
continue for any kind of write error, but perhaps st(4) is not
returning the error dump expects at eom?

>How-To-Repeat:

As above.

>Fix:

trying pax instead
>Release-Note:
>Audit-Trail:
>Unformatted: