Description of problem: When attempting to install using the anaconda pacakge from 20070523 which supposedly fixes this problem I still get a backtrace when installing on my mdraid machine. Version-Release number of selected component (if applicable): anaconda-11.2.0.63-2 How reproducible: 100% Steps to Reproduce: 1. boot boot.iso 2. set http mirror 3. wait Actual results: When X comes up and you click next, anaconda elects to throw a backtrace at you Expected results: Additional info:
Created attachment 155310 [details] anaconda dump
Okay, so.. you've got dmraid on.. sata_nv? Can you give us a bit more info about the RAID set up - number / type of disks, controller, etc?
mdraid not dmraid Woods. it's an nforce3 SATA controller with dmraid capabilities turned off (since historically Linux has sucked on that, I plan to change it back for F8 to break it in interesting ways). 2 WDC WD4000YR-01P harddrives /boot on RAID1 /, /home and SWAP on LVM on top of a RAID0
ah, sorry.. it's late and I'm mixing up mdraid and dmraid again. according to pjones there was a bug in the mdraid metadata finding stuff that should be fixed in anaconda 11.2.0.64, so hopefully the next rawhide should be better.
I managed to pull off an mdraid upgrade yesterday without a problem, whereas I couldn't a few days ago, so at least we're closer...
We're trying to start the device md-1 which is weird. What's the output of mdadm -E for /dev/sda2, /dev/sdb1 and /dev/sdb2?
Just set up a box with a pair of WD 120G ATA drives, 100M software raid1 /boot, the rest software raid0 with lvm atop it, featuring /, /home and swap. No more explodeyness, only oddity was grub needing to be reinstalled rather than upgrading the existing config for some reason.
Works for me as of F7 RC2.