Subject: Archive 4586NP woes
To: None <current-users@NetBSD.ORG, mail2news@replay.com>
From: VaX#n8 <vax@linkdead.paranoia.com>
List: current-users
Date: 07/23/1998 02:47:30
Tape Drive Questions:

    Drive: Archive/Python/Seagate 4586NP 4mm tape changer

    I have been noticing some strange behavior.

    First, the documentation I was given (in the form of some
    photocopied sheets) is rather lacking.  In fact, the SCSI ID
    jumper settings were left-right reversed.  This cost me quite
    a bit of time to figure out.

    After the drive writes to a tape, it won't respond
    to the Open/Close button (or the Step button, for that matter)
    until the tape has been ejected, even if the tape is rewound.
    Is this normal?

    On various occasions, I have witnessed the green "cassette" LED
    to blink slowly, or the amber "drive" LED to blink quickly.
    What do these mean?

Hardware or OS questions:

    Controller: based on NCR 53c810
    OS: a recent NetBSD (>= 1.3)

    Upon rewind, the drive retains exclusive access to the SCSI bus.
    I can think of three explanations:
        1) The drive does not support disconnect/reconnect.
        2) The controller does not support disconnect/reconnect.
        3) The OS does not support disconnect/reconnect.
    Any guesses?

    Furthermore, if the "offline" command is given, and the drive
    does a long rewind, unload, reloads the next tape, and there
    are significant accesses to the SCSI bus during this time,
    the machine does a hard reboot.

    Typical system logs are:

... ncr0: timeout ccb=0xf0874000 (skip)
... st1(ncr0:6:0): non-media hardware failure, data = 00 00 00 00 44 bc 00 3c 00 00
... st1(ncr0:6:0): not ready, data = 00 00 00 00 00 00 00 00 00 00

    The timeout is by far the most frequent.

OS question:

    I also had problems with the machine logging a line mentioning
    a timeout, and the CCB being dequeued.  This frequently led to
    a hung SCSI bus.  Thoughts?
--
Paranoia is underrated.