home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!cf-cm!news
- From: spxpmf@thor.cf.ac.uk (Phillip Fayers)
- Newsgroups: comp.sys.sun.admin
- Subject: Re: ATTENTION: A problem with dump under 4.1.1
- Keywords: tape, cleaning
- Message-ID: <26109.9207241157@thor.cf.ac.uk>
- Date: 24 Jul 92 11:57:11 GMT
- References: <HFL-H1A7T@linac.fnal.gov> <1992Jul23.132034.6025@csc.canterbury.ac.nz> <15732.9207240920@thor.cf.ac.uk>
- Sender: news@cm.cf.ac.uk (Network News System)
- Organization: Physics Dept. University of Wales, Cardiff
- Lines: 50
- X-Mailer: Cardiff Computing Maths PP Mail Open News Gateway
-
- In article <15732.9207240920@thor.cf.ac.uk> spxpmf@thor.cf.ac.uk (Phillip Fayers) writes:
- |In article <1992Jul23.132034.6025@csc.canterbury.ac.nz> agw@math.canterbury.ac.nz (Allen Witt) writes:
- ||
- || billq@fnal.gov (William R. Quayle) write
- ||>In article <1992Jul21.195006.112389@zeus.calpoly.edu>, etsiao@joule (Eddie Tsiao) writes:
- ||>|> In article <1992Jul20.073110.20195@aristo.tau.ac.il> shani@GENIUS.TAU.AC.IL (Oren Shani) writes:
- ||>|> We have had (are having?) similar problems. We occationally get an error:
- ||>|>
- ||>|> DUMP: Tape write error 753 feet into tape 1
- ||>|> DUMP: fopen on /dev/tty fails
- ||>|> DUMP: The ENTIRE dump is aborted.
- ||>|>
- |I've been having very simlar problems, last nights dump failed with-
- | DUMP: dumping (Pass IV) [regular files]
- | DUMP: Tape write error 10 feet into tape 1
- | DUMP: fopen on /dev/tty fails
- | DUMP: The ENTIRE dump is aborted.
- |
- |I've had this happen before (oddly enough its usually 10 feet into the
- |tape) so I usually redo the dump by hand when I arrive in the morning
- |and it works! I've tried cleaning the drive but it doesn't seem to do
- |anything.
- |I'm about to rewrite the dump scripts so that they reposition the tape
- |and have another go if they fail.
- |I also see occasional errors logged to the console:
- |
- |Jul 24 03:05:42 orion vmunix: st1: I/O request timeout
- |Jul 24 03:05:42 orion vmunix: sm0: target 5 reverting to async operation
- |Jul 24 03:05:42 orion vmunix: sm0: resetting scsi bus
- |.......
- |
- |although this last one is the first to actually coincide with a tape
- |error.
- |I'll probably contact our tape drive supplier to see if he has any
- |clues.
- Well I just got off the phone, the answer from our suppliers technical
- department is that the Exabyte 8500 are not supported fully under SunOS
- 4.1.1 (and yes he knows about the hack to stdef.h to enable us to use
- different densities). It seems that the Exabyte uses a SCSI-2 interface
- which, in certain circumstance, does a SCSI disconnect when its busy, so
- it frees up the SCSI bus for other devices. As SunOS 4.1.1 doesn't
- understand this disconnect it sometimes causes a SCSI reset, it seems
- that at other times its just causing a tape error. It seems that if you
- don't want problems upgrade to 4.1.2, unfortunately I don't think that
- we can.
-
- --
-
- Phillip Fayers email: fayers@cardiff.ac.uk
- Sun admin/support/programming phone: 0222 874000 x 5282 (UK)
-