Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!ut-emx!margo.ots.utexas.edu!jah
From: jah@margo.ots.utexas.edu (Jeff Hayward)
Newsgroups: comp.unix.ultrix
Subject: Re: DS3100, Ultrix4.1, Exabyte, and dd
Summary: WangDAT 1300 also
Keywords: multiple files with dd
Message-ID: <47879@ut-emx.uucp>
Date: 25 Apr 91 15:13:50 GMT
References: <1991Apr22.223308.5674@colorado.edu> <1991Apr23.133952.17133@mercury.cair.du.edu>
Sender: news@ut-emx.uucp
Organization: The University of Texas
Lines: 31


In article <1991Apr23.133952.17133@mercury.cair.du.edu> mbrookov@diana.cair.du.edu (Matthew B. Brookover) writes:
|In article <1991Apr22.223308.5674@colorado.edu> devin@boulder.Colorado.EDU writes:
|>apparently using the dd files= option causes problems...All of the requested
|>files are read off of the tape except for a few mangled bytes right around
|>the EOF markers (between the files).  There are obvious workarounds, but
|>we would like to be able to do this.  Has anyone else seen this problem
|>and/or a fix for it?  Thanks.
|
|Yes, I have been having this problem.  What computer and tape drive are
|you useing?  I have a DECsystem 5400 running Ultrix 4.1 and a Micro
|Technology 8mm tape drive (Exabyte) with a QTS30 controller.
|
|I do not even have to write multiple files, just 1 file is enough to
|screw up the last few bytes of the file.

I've been having a similar problem with a WangDAT 1300 (4mm DAT drive)
under Ultrix 4.0 on a DS3100.  When dd reads the last block, usually a
partial block, the read() returns -1 and errno is 5.  The data
returned in the read buffer appears to be fine, but is discarded.

Another oddity of the ultrix/WangDAT combination is that the first
open() done on the device after a new tape is inserted returns an
error.  I presume this has to do with the device passing back the SCSI
status indicating a new tape is present.

Anyone know what's going on here?
-- 
Jeff Hayward
The University of Texas System          +1 512 471 2444
Office of Telecommunication Services    jeff@nic.the.net