Serious issue with booting after updates

[vinmansbrew]

I ran into a super serious issue after running yum updates. the server boots, but I get nothing, just a blank screen with a -, nothing more. No grub, nothing. I cannot type at the prompt or anything. I have no idea what happened. There was a kernel update, but only 1.
I have tried to reinstall grub, no joy. I can boot to recovery off an iso. BTW, this is a virtual machine. I did a few servers tonight, the rest went through without a hitch.
Any one on with an idea? Please help!

[CaViCcHi]

boot from live cd, mount your /boot partition and take a look at your grub file

make sure that the default's index matches an actual kernel that exists under boot

[vinmansbrew]

well the grub shows kernel /vmlinuz-2.6.32-696.el6.x86_64
rpm -qa shows kernel-2.6.32-696.1.1.el6.x86_64

Would that be correct?

[vinmansbrew]

I did try doing an rpm uninstall of the kernel, didn't seem to work all the way. I can get a network connection working, so maybe a yum uninstall of the kernel would be an option?

[CaViCcHi]

just have the disk mounted rw and modify the text into the name of the physical file on disk and it should boot fine

[TrevorH]

If you don't have a vmlinuz-2.6.32-696.1.1 yet yum says that kernel is installed then it sounds like you have a disk problem. Most likely thing I can think of is that you are using software RAID 1 for your /boot and the array is broken and not replicating to the drive that the BIOS is trying to boot from.

[vinmansbrew]

Well, none of the other servers on this vm host are running into issues, and since it happened after the updates, I'm fairly certain it is something os related. I may be wrong, but I don't think so.

CaViCcHi, which file are you referring to? I mean, I can chroot fine, and I have been able to do whatever I want to the file on the drive.

[CaViCcHi]

for example the file

/boot/efi/EFI/redhat/grub.conf

or

/boot/grub/grub.conf

or whatever grub.conf under /boot (depends on your configuration, but you really only should have one)

then it should be somewhat like this

Code: Select all

# This is the default, 0 means the first one, 1 the second one... 
default=0

# First one is index=0. Don't touch the root (hd0,0) or whatever yours says
title Your CentOS 6 (2.6.32-642.6.2.el6.i686)
    root (hd0,0)
    kernel /vmlinuz-2.6.32-642.6.2.el6.i686 ro ... YOUR KERNEL LINE... rhgb quiet
    initrd /initramfs-2.6.32-642.6.2.el6.i686.img
    
# If I had a second one it'd be index=1 and so on

from what I understand you should have something in there that says

grub.conf

Code: Select all

default=NOTSURE

title Your CentOS 6 (2.6.32-696.el6.x86_6)
    root (hdX,Y)
    kernel /vmlinuz-2.6.32-696.el6.x86_64 ro ... YOUR KERNEL LINE... rhgb quiet
    initrd /initramfs-2.6.32-696.el6.x86_64.img

and on disk you have

Code: Select all

/boot/vmlinuz-2.6.32-696.1.1.el6.x86_64
/boot/initramfs-2.6.32-696.1.1.el6.x86_64.img

if this is the case, then change the grub.conf into

Code: Select all

# Change this to 0
default=0

# Bring this one to the first one in list 
title Your CentOS 6 (2.6.32-642.6.2.el6.i686)
    root (hdX,Y)
    kernel /vmlinuz-2.6.32-696.1.1.el6.x86_64 ro ... YOUR KERNEL LINE... rhgb quiet
    initrd /initramfs-2.6.32-696.1.1.el6.x86_64.img

basically all you do is add ".1.1" if this is the case...

otherwise all you have to do is match the file name (both initramfs and vmlinuz) you have under /boot/

[vinmansbrew]

Well, thanks for the help. I found what I was doing. I was doing grub-install /dev/sda1, and it should have been /dev/sda.
Very simple mistake, but I just didn't know. I have, however, learned quite a bit more with this fiasco!
Thanks for the help, CaViCcHi!

[CaViCcHi]

there's nothing better than making mistakes to learn (it's just better if in the process you don't lose data as well eheh)