From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.0.12) Gecko/20071019 Fedora/1.5.0.12-5.fc6 Firefox/1.5.0.12 Description of problem: Kickstart install via PXE. System uses nVidia nForce4 SLI chipset. Reading the traceback file, it looks like anaconda isn't able to write the MBR. Same system has worked with FC6 (and FC5). Version-Release number of selected component (if applicable): anaconda 11.3.0.50 How reproducible: Always Steps to Reproduce: 1. Install f8 using PXE or DVD (always used same http:// install tree) 2. Usually partition drives as dm raid 1, but fails regardless 3. Actual Results: Same traceback regardless of test conditions. Tried booting other media, not using RAID, wiiping the MBR from the two drives in question, swaping the drive boot order, resetting the machine's BIOS back to default values... Expected Results: Grub should have installed to the primary drive's mbr, and not generated a fatal exception. Additional info: Other nForce4 SLI machines can install from this same installation tree, using the same kickstart file.
Created attachment 289734 [details] traceback file Here's what anaconda copied
Created attachment 289735 [details] Kickstart file Just for fun, here's my kickstart file
In the ks file, all of the partitioning information is commented out and it looks like a "real" ks file as opposed to one that anaconda spit out. Is this an oversight?
On futher digging, this is just "stupid user error" The kickstart file is "real" and the line: bootloader --location=mbr --driveorder=sda --append="rhgb quiet" is what was causing my problems. If I omit the "--driveorder=sda" parameter, everything works perfectly. Sorry for the spurious report, and kudos on your fast response to it! :-)