Topic: server hangs on boot after update

[jimgoode]

Hardware:
Dell Optiplex GX200 (shipped Feb. 2001)
Intel P3 733 MHz (Coppermine)
256 MB RAM
2 x 10 GB IDE HDD - RAID 1

I am testing the upgrade process on a test system before attempting it on production. I:
Installed SME 6.0.1-01
Upgraded to SME 7.4
Applied all software updates and the server hung during the boot process displaying only 'LI' on the monitor

I duplicated the process and on the second attempt omitted the software updates for kernel-smp.i686 2.6.9-78.0.13.EL, kernel.i686 2.6.9-78.0.13.EL,
kmod-appletalk.i686 2.6.9-78.0.13.EL, kmod-ppp.i686 2.6.9-78.0.13.EL,
kmod-slip.i686 2.6.9-78.0.13.EL, openssl.i686 0.9.7ª-43.17.el4_7.2
and the system appears to be fine. The original kernel was 2.6.9-78.0.8.EL.

Based on my hardware, can anyone tell me if it is incompatible with the newer kernel updates? If yes, how can I tell in advance if I will encounter the same problem with the production server?

Production hardware:
Generic tower - Elite motherboard
AMD 2500+ CPU (1.1+ MHz)
1 GB RAM
2 x 300 GB IDE HDD - RAID 1

Thanks,
Jim

[David Harper]

Welcome to the forums

Please file a bug report so this issue can be investigated. However, as this is likely to be a kernel issue it is possible that the problem originated upstream, either with CentOS or Red Hat. You could confirm this by installing CentOS 4.7 and seeing how the machine fares.

With regards to your other server, chances are everything will be fine. But I would make a backup before doing the upgrade, just in case.

[jimgoode]

Thank you, David.

I'm going to download the CentOS 4.7 and give it a try like you suggest. Since that will wipe out what I've done so far, I decided to apply those 6 omitted packages and see what happens. Same result, but I noticed that the kernel and kernel-smp packages both gave an error "grubby: unable to open /boot/boot.b no file found". Sure enough, that file isn't there. The boot hangs with "LI" displayed on the monitor.

I rebooted with the SME 7.4 CD and typed SME rescue. I can see in the /mnt/sysimage/boot directory that both sets of kernel files are there and grub.conf has the default set to the first one loaded with SME 7.4 (2.6.9-78.0.8.EL).

Is there a way to salvage the system and get past the hang up?

Thanks,
Jim

[David Harper]

Possibly. Before you wipe anything I would definitely file the bug, so the developers have a chance to walk you through any troubleshooting or recovery steps that might be possible.

When you've filed the bug, please post a link to it in this thread for future reference.

[jimgoode]

Bug 5088 submitted.