From Bugzilla Helper: User-Agent: Mozilla/4.77 [en] (Win95; U) Description of problem: Dell 4400, Raid Controller 3/Di, delivered with Redhat 6.2 (worked great). Attempted to upgrade to 7.1 (SeaWolf), Pro Upgrade seemed to be fine, but all subsequent attempt to boot fail with kernel panic: request_module[block-manager-8] Root fs not found VFS cannot open root device "806" or "08:06" please append a correct "root=" boot option Am able to boot from the install disk in rescue mode. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. Turn on the machine, let it try to boot. 2. 3. Actual Results: Boot process starts. Raid controller starts. Kernel is loaded and starts executing execution ends with kernel panic and previous output Expected Results: Boot process completes. Additional info: Have attempted 2 1/2 months of fixes on the customer service (case 194547). They've sent me here. These are a summary of my responses to their requests: e2label showed that the labels had been stripped from all of the partitions. Reset all of the appropriate labels. Same result on reboot attempt: From rescue boot there is no /etc/fstab. In fact /etc is almost empty, e.g.: group, protocols, services are all links to /mnt/runtime/etc. All files therein have dates of <today>, or April 09 03:38. Furthermore, the entirety of / also appears to be sparcely populated. Tracked down that the previous image is mounted at /mnt/sysimage, all of the filesystems mounted there appear good (/mnt/sysimage/etc/fstab is correct). After Booting into rescue mode.... The only thing in /var is the directory "state." The earliest dmesg and messages files that I can find on the system (find / -name xxx) are in /mnt/sysimage/var/log, and are from Nov 04 (the last good boot prior to the upgrade). Neither lilo.conf nor fstab files are in /etc (there are only 11 files in /etc). I have executed 'lilo -C /mnt/sysimage/etc/lilo.conf', to confirm that the included version is the one in use; have also tried removing the 'linear' directive (as it is not in lilo.old). Same kernal panic in both cases. mkinitrd returns: /mnt/sysimage/sbin/mkinitrd: No such file or directory. There are no initrd files for the new kernel (although they do exist for the earlier [6.2] version that came installed on the machine). I was unable to get a man page for mkinitrd. Perhaps if you could give me the parameters, we might be able to get that to run. When I tried it before, all of the partitions were still mounted at /mnt/sysimage -- since the fstab file references them all as mounted from / boot into 7.1 rescue mode chroot /mnt/sysimage mount -a /sbin/mkinitrd /boot/initrd-<version>.img starts working, but returns with returnes: "no module percraid found for kernel < kernel-version > That's when customer service stopped responding.
percraid Hmm... that's the OLD name for "aacraid".... No wonder it doesn't work if it tries to load a non-existant driver. Looks like you're almost there. Just before typing the mkinitrd in what you described last, edit (with vi or whatever editor you like) the /etc/modules.conf file and change "percraid" into "aacraid" and then generate the initrd as described, or rather as mkinitrd /boot/initrd-2.4.2-2.img 2.4.2-2 (eg repreate the version as second argument. If you use the SMP kernel add smp etc etc)
Also had to update /etc/lilo.conf to add initrd lines to both kernals. After that, everything was golden. Thanks, nice to have the machine back.
Can you post the output of 'lspci' on the system in question?
closed, no response.