I've been assigned the task of figuring out why one of our recently configured advertising servers with fedora core 3 crashes. The system was initially configured with Core 1, and upgraded to core 3 with Yum. Looking through the system log revealed messages along the lines of kernel: kfree_skb on hard irq errors. These messages were popping off by the minute, and messages was over 3 gigs in size. Interestingly enough the kernel was of the 2.6 variety yet /etc/redhat-release said the system was still core 1. Somehow during the core 1 to core 3 upgrade everything wasn't completed.
After much time spent on the net getting yum to work again (many rpms weren't the correct version to support the yum rpm on the box) I was able to upgrade the box to Core 3. The system comes up on the new kernel (2.6.11-1.14_FC3) but locks up usually a couple times a day. The kfree_skb messages are gone from the system log, bt nothng is there in regards to what's causing the lockups.
I've been trying to put a custom kernel on the box in order to troubleshoot, however I've been having problems getting past the mounting of the rootfs
Messages I've seen so far are below
VFS: can not open root device "LABEL=/" or unknown-block(0,0)
Please append a correct "ROOT=" boot option
kernel panic - not syncing :VFS: unable to mount root fs on
I searched around the net and found some information regarding some compile options that I may have left out. I added msdos filesystem support, ext3 support(even though we have changed it to ext2 in fstab), and XFS support
After all of that I'm getting messages along the lines of
mount: error 22 mounting ext3
mount: error 2 mounting none
Switching to new root
switchroot: mount failed: 22
umount /initrd/dev failed: 2
Kernel panic - not suncing : attempted to kill init!
I'm thinking there's a problem in the grub configuration. I've since modified the grub entry to point to the partition root is on (/dev/hdb8) instead of root=LABEL=/
I've also changed the kernel line to include an initial ramdisk that I made using the mkinitrd command against the 184.108.40.206 kernel bzImage I made from the kernel sources at kernel.org.
I'm hoping the box will boot now, but I can't be sure, and am looking for advice. The system is a customers and it's not local to me, so any failed boot is frustrating. Please provide insight if you have any. Thanks.