jaysun@omni.eng.clemson.edu (Jay Williamson) (11/07/89)
We have a couple Sun 3/80's with floopy disks. The problem is that fdformat will not work, neither will eject. It seems to be some problem with the device names. We did a MAKEDEV fd0, just like the manual says (The manual says this mounts the device I was not aware that doing mknod mounts anything). This created /dev/fd0[a-h] and /dev/rfd0[a-h]. If eject is typed then it gives no error, it returns like it did the right thing but does not eject the disk. If you try an fdformat -f as anything but root then it returns "fdformat: could not open "/dev/rfd0c": Device busy". If root trys it then it returns Format /dev/rfd0c, 80 tracks, 18 sector/track ... Format failed : No such device or address MAKEDEV looks reasonable but is it leaving something out? The kernel is definitly finding the device at boot time but we are running a strange configuration is it matters. Only / and /tmp are local. The rest of the local space is swap. If this has been discussed and dropped please excuse me and point me in the right direction. If it hasn't then send answers to me and I will summarize later. Jay Williamson Clemson University Systems Programmer Computer Science Dept.
harv@herkimer (Harvard Townsend) (11/21/89)
In article <2872@brazos.Rice.edu> jaysun@omni.eng.clemson.edu (Jay Williamson) writes:
]>X-Sun-Spots-Digest: Volume 8, Issue 189, message 15 of 22
]>We have a couple Sun 3/80's with floopy disks. The problem is that
]>fdformat will not work, neither will eject. It seems to be some problem
]>with the device names. We did a MAKEDEV fd0, just like the manual says
]>(The manual says this mounts the device I was not aware that doing mknod
]>mounts anything). This created /dev/fd0[a-h] and /dev/rfd0[a-h]. If
]>eject is typed then it gives no error, it returns like it did the right
]>thing but does not eject the disk. If you try an fdformat -f as anything
]>but root then it returns "fdformat: could not open "/dev/rfd0c": Device
]>busy". If root trys it then it returns
]>
]> Format failed : No such device or address
]>
]>MAKEDEV looks reasonable but is it leaving something out? The kernel is
]>definitly finding the device at boot time but we are running a strange
]>configuration is it matters. Only / and /tmp are local. The rest of the
]>local space is swap. If this has been discussed and dropped please excuse
]>me and point me in the right direction. If it hasn't then send answers to
]>me and I will summarize later.
We had the same problem with our 3/80s. It turned out that the addressing
switch on the drive itself was set to be device 1 instead of 0 (it is a
small black 4-position slide switch on the side of the drive -- it was set
in the 2nd position rather than the first). Setting the switch to 0 cured
the problem. We could not find this documented anywhere. My guess is
that this was set to 1 rather than 0 at the factory because the majority
used to be shipped as an external drive rather than internal. Who knows.
Anyway, it works now. Give it a try (our area Sun technician hadn't run
into this either).