GRUB problem encountered with 6.9 upgrade

[Bruce]

Yesterday I updated a file server from 6.8 to 6.9. The install appeared to be clean, and I decided to reboot when it was done. The server made it through all the BIOS stages, but then it hung with a flashing cursor on a blank screen.

I don't see anyone else reporting problems with this update so there may be something unusual with my hardware or filesystem, but I thought I should report what happened with this server in case it is helpful for someone.

I downloaded an ISO of CentOS 6.9 minimal, burned a CD-ROM and rebooted the server. I used rescue mode, and after looking around at the filesystem for a while I was reasonably confident the data was intact, so I re-installed GRUB:

Code: Select all

chroot /mnt/sysimage
/sbin/grub-install /dev/md0

After a reboot, the server spent some time updating selinux attributes (I think it was), rebooted itself again, and came up fully functional.

My boot disk uses software RAID1, which is what the /dev/md0 is about. Two drives /dev/sda and /dev/sdb, raided and partitioned for /, /sboot, /tmp and swap. There's a couple other drives in the system as well, but they only hold data and aren't part of this RAID array.

[CaViCcHi]

wait your /dev/sdX1, where you keep your EFI/grub/boot information basically, is on raid?

by standard the grub.conf information should be OUTSIDE of the raid array when running with software raid. and your fstab should point directly at one of your disks (unraided, if that's a word )

I think it's possible that your grub info got garbled on one of your disks.