Red Hat Bugzilla – Bug 38645
Anaconda fails to run after loading symbios drivers
Last modified: 2007-04-18 12:32:57 EDT
From Bugzilla Helper: User-Agent: Mozilla/4.72 [en] (X11; U; Linux 2.2.14-5.0smp i686) Attempted to run RH7.1 installer from CDROM using following flags: linux expert text noprobe nousb (and various permutations therein). After finding and loading the symbios driver (ncr version), it hangs on the load of ext3.o Running same installer with additional noapic flag creates a failed load of the glibc.so.6, again, dies nicely, unmounts drives, and asks to be rebooted. Reproducible: Always Steps to Reproduce: 1. Run 7.1 installer from cdrom 2. Manually select the SCSI driver 3. Watch it fail over and over again Actual Results: On an Alt-F3 screen, it shows the various driver loads, and fails on ext3.o On the Alt-F1 screen, anaconda fails, dismounts the drives nicely, and tells me its ok to reboot. Expected Results: The installer should run...
Created attachment 16987 [details] Output of lspci -vb
Created attachment 16988 [details] Output of lspci -vn
More detail: Actual verbatim output begins with: Running anaconda...install exited abnormally -- received signal 11 Sending termination signal...done Sending kill signal...done Unmounting... <snip> You may now reboot your machine
Looks like a kernel problem...changing component
Is this a IDE cdrom driver ? If so, please try using the "ide=nodma" option in the place where you did "noapic" before. If that doesn't fix it, it's an anaconda segfault
That ('ide=nodma') was it. Thanks much! Strikes me as kinda odd that support had me report it as a bug, tho... Shouldn't this be something that belongs in the knowledgebase (e.g., 'Installing to SCSI Hard-Drives with IDE CD-Rom controllers'). Just wondering. So do I close the bug or do you?
Support has a kb article about IDE cdrom drives. Maybe the relation to SCSI messed the "ah it must be that" up. Can you do me a favour and do a cat /proc/ide/hda/model (replace "hda" with the real device of your cdrom drive) ? This way, we can mark your cdromdrive as non-DMA capable and fix it for future versions. Thanks.
cat /proc/ide/hda/model shows: HITACHI CDR-8435
Thank you for the information. I've added it to the list of cdrom drives the kernel will not use DMA on. Two days ago I added "HITACHI CDR-8335" to the list.. looks like they made a series of them ;)