malc@uunet.uu.net (Malcolm Carlock) (06/05/91)
We have a Sun IPC with attached CDROM drive which works up until after
the second or third eject operation, after which time attempts to mount
the CDROM filesystem (even from disks that had been successfully mounted
a few minutes before) fail with "I/O error". The following methods
consistently work/don't work in clearing up the problem, which has been
present since the drive came out of the box:
Cycle power on CDROM drive doesn't work
"shutdown -h" CPU, reboot, no power-off works
Cycle power on CPU, reboot works
Cycle power on entire setup works
It looks like the kernel and/or some hardware in the CPU are getting
angry about something that only a kernel reboot can fix. Has anyone
else run into this problem? Does anyone know what sort of signal/ioctl
operation might clear things up without rebooting?
Thanks in advance,
Malcolm L. Carlock Internet: malc@unr.edu
UUCP: unr!malc
BITNET: malc@equinoxrmilner@zia.aoc.nrao.edu (Ruth Milner) (06/11/91)
In article <2941@brchh104.bnr.ca> unr!malc@uunet.uu.net (Malcolm Carlock) writes: >We have a Sun IPC with attached CDROM drive which works up until after >the second or third eject operation, after which time attempts to mount >the CDROM filesystem (even from disks that had been successfully mounted >a few minutes before) fail with "I/O error". The following methods >consistently work/don't work in clearing up the problem, which has been >present since the drive came out of the box: We have been having a very similar problem with a CD-ROM attached to a SPARC2 (but we moved it to a SPARC1 and it failed there too). It will mount a couple of times and then start failing. The error we get from the mount is "no such device or address", and when you look at /usr/adm/messages it says sr0: disk not responding to selection sr0: Unit not Ready. Additional sense code 0. Please don't suggest that we are missing some kernel support; sr0 is there, HSFS is there, the SCSI address and termination are correct, etc. etc. It does work, just not for any length of time. There are an internal disk, an external disk, and an Exabyte also on the bus. We first notified Sun of this in January. Both the drive and the SCSI cable have been replaced to no avail. As noted above, the problem happened on two CPUs, so it seems unlikely to be a CPU problem. It was also tried while the SPARC1 was at 4.1 (the SS2 has of course always been at 4.1.1). Our local Sun service rep. has been able to produce a problem, though not the same one, by mounting and unmounting a non-HSFS CD without the -r option (HSFS won't let you do this). On the umount, the system attempts to write to the disk and generates several SCSI errors. On about the third time, the umount hangs the system. It looks as though there are indeed some peculiar problems with the Sun CD-ROM. Ruth Milner Systems Manager NRAO/VLA Socorro NM Computing Division Head rmilner@zia.aoc.nrao.edu