From owner-freebsd-questions Wed Apr 10 07:17:30 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id HAA09363 for questions-outgoing; Wed, 10 Apr 1996 07:17:30 -0700 (PDT) Received: from mail.nyc.pipeline.com (root@mail.nyc.pipeline.com [198.80.32.13]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id HAA09358 for ; Wed, 10 Apr 1996 07:17:27 -0700 (PDT) Received: from pipe4.nyc.pipeline.com (axon@pipe4.nyc.pipeline.com [198.80.32.44]) by mail.nyc.pipeline.com (8.7.3/8.7.3) with ESMTP id KAA24259 for ; Wed, 10 Apr 1996 10:17:10 -0400 (EDT) From: "Amir Y. Rosenblatt" Received: (axon@localhost) by pipe4.nyc.pipeline.com (8.6.9/8.6.9) id KAA03610 for freebsd-questions@freebsd.org; Wed, 10 Apr 1996 10:17:05 -0400 Message-Id: <199604101417.KAA03610@pipe4.nyc.pipeline.com> Subject: tar and restore problem To: freebsd-questions@freebsd.org Date: Wed, 10 Apr 1996 10:17:02 -0400 (EDT) X-Mailer: ELM [version 2.4 PL22] Content-Type: text Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk After bringing my machine up and running it for the better part of a week, I decided that I needed to reconfigure the disk partitions. So I did a set of filesystem dumps to a 4mm DAT, but when I tried to do test restores of them, I got an error message ("Input/output error"). I tried about a dozen different sets of parameters but none of them, let me do a restore. So, I tried using tar instead (I did my initial installation from tape, as well as having installed a bunch of the initial software that way). I did some test tart and untars and they worked fine. So, I crossed my fingers and did a "tar -cvf /dev/rst0 /" and 15 minutes later it finished. I then went and reinstalled the machine from the 2.1 install media (the same tape I used the first time) and prepared to restore the OS to the way it had been before the reconfig. Howerver, when I attempted to untar it (using "tar -xvf /dev/rst0") I got the following error: st0: 65536-byte record too big tar: read error on /dev/rst0: Input/output error Then I got two console messages, both reading as follows: /kernel: st0 65536-byte record too big I attempted to list the contents of the tape, using "tar -tvf /dev/rst0" and that worked fine. But I was still unable to untar the contents of the tape. OS-wise, when I did the backup I was running a 2.1R kernel on a -stable OS (I'd had some PPP problems with my -stable kernel which I had planned to deal with tonight and there had been no problems with the setup up to that point). The tape drive is an HP SureStore SCSI-2 8 gig 4mm DAT drive. My hard drives are Seagate ST32550W Baraccudah 2LPs (I have 2 of them). I'm using an Adaptec 2940W as my SCSI-2 host adapter. The machine itself is a P120, with an Asus motherboard, 32 meg of RAM and a Mach64 graphics card. I recompiled my kernel hoping that might help (couldn't hurt, as it brought it closer to the way I'd had the kernel set up originally), but there are still some differences between my current kernel and the original one. The original one had the SCSI drives hardwired to their device files (i.e. SCSI device 0 was hardwired to sd0, etc). Any suggestions? -Amir