From Bugzilla Helper: User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.4.3-2.9.2smp i686) Description of problem: I have a 100% scsi system, Adeptec 29160, IBM scsi U160 disks, Yamaha scsi cdrw. The RC2 kernel detects what I believe to be false media errors on the scsi cdreader. I cannot install. I have written disc1 2 times, and each time, a media error is detected in one package, or another during installation, and retry will not read the block/package. I can reboot to an earlier RH7.1 installation, and I can read the same cdrom on the same yamaha drive, without a single error. This is reminsicent of the aha7xxx errors we used to have during the early days of 7.1 beta. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. set up a scsi only system 2. install from a yamaha scsi cdrw (same one that wrote the cd's - this REALLY SHOULD work !!!) 3. Actual Results: failed install. media error detected, and once detected cannot continue install. retry does no good as it sees the false media error again and again, on an otherwise idle system. Expected Results: installed system. Additional info: can't install on aic7xxx scsi based systems - I set this as high priority because of that.
Switching to kernel since that's where the problem would lie
Is this by accident a 440GX chipset system ? If so, does passing "apic" on the syslinux prompt help ?
This a SMP system, so apic is already enabled. The chipset is VIA.
hmm. I just thought about this again. While it is true that my box is SMP/Via chipset, it just occured to me that when I'm booting the installation cd, it's booting a UP kernel, so apic is probably not being set. Let me try enabling this and see if this helps.
still fails. I tried lowres apic at the boot prompt (I need "lowres" for my graphic card support). I still get a false media error, but at a different place.
it looks like this might be hardware. I swapped the cdrom drive, used the same media (created on the original cdrom drive), and the install went fine. My big clue was when a cp -a of the cdrom got i/o errors... I'm closing this as NOTABUG -Greg