Red Hat Bugzilla – Bug 54756
Cannot boot after using up2date to update kernel
Last modified: 2015-01-07 18:52:10 EST
From Bugzilla Helper:
User-Agent: Mozilla/4.0 (compatible; MSIE 5.01; Windows NT 5.0)
Description of problem:
First I ran up2date (version 2.5.4-x) as a GUI to update from 2.4.2 to the
latest version. This downloaded the .rpm file but failed with a message I
didn't think to write down. It had something to do with being unable to
verify header information. Since things often work better from the
command prompt, I ran "up2date kernel" from a terminal window and all
indications were that it completed successfully.
I checked lilo.conf and the new kernel information was in place and
identical to the old information (except for the image= tag). To make
sure, I ran lilo.
However, when I rebooted, the process displayed "LIL-" and halted. (I'm
still working on getting my system back up.) Note that this is a multi-
boot system with the /boot directory and lilo installed on a separate
partition (/dev/hde2). The other operating systems continue to boot
I tried changing lilo to boot the old kernel but, of course, this failed
since the boot process never gets that far.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Run the command "up2date kernel"
Actual Results: System fails to boot.
Expected Results: I rather expected the system to boot with the new
Adding to the previous information, I also ran "lilo -u" which, I understand,
should uninstall the boot sector and replace it with the old one. The
following messages were displayed:
"Reading boot sector from /dev/hde2"
"Fatal: Boot sector of /dev/hde2 doesn't have a LILO signature."
That interested me since I had used lilo to install it.
I also should note that I was a little hasty to say that my other operating
systems still booted. The Windows 2000 Server installation in the adjacent
partition kept repeating the login screen.
Since I hadn't placed any important information on the drive, I adopted the
draconian measure of wiping and repartioning the drive. Now I'm re-installing
Note: I just was told that I should have run "fdisk /mbr" and then reinstalled
the NT loader. I don't know if that would have fixed the "LIL-" problem. BTW,
there seems to be no information about correcting a "LIL-" problem; only the
information that this means that the "descriptor table" is "corrupt" without
saying what the descriptor table is or how to uncorrupt it.
Hmm. This sounds like it's probabaly actually a lilo bug. The up2date
client will change the config file, and run lilo in test mode, and
if succesful, run lilo in live mode.
The fact that running lilo by itself seems to indicate that the config
file was valid. It is possible that the config file was valid, but
incorrect, but I havent really seen any other reports of this.
Configuration errors generally are found when installing the bootloader.
I dont suppose you have access to the lilo.conf that was created and installed?
I'm not sure if I still have that lilo.conf file. I had to repartition my
system and do a clean install. I'll check to see if I made a copy. It looked
the same as any other lilo.conf I've seen.
I'm afraid that I didn't keep a copy of lilo.conf before wiping my system
(again). I do remember examining it carefully and finding nothing unusual.
BTW, although you said you hadn't seen this problem before, I notice that in
the user forums there's at least one message from a person who had the exact
same problem after using up2date to update his kernel.
I belive most of the kernel/bootloader related issues should
be fixed in current versions of the client so I'm closing this
If that is not the case, please reopen.