Help after kernel update failed. EL- ELsmp

[blkjck]

I let yum run an update the other day and it pulled down 2 files for a new kernel. 2.6.9-55.0.12.EL and 2.6.9-55.0.12.ELsmp This was the first kernel update I have done with yum and it did not go well. Upon reboot the server never came back up and no powercycle helped. I found the post about editing the grub.conf to modify it next time with --once entry. Lesson learned.
So someone at the datacenter went in and got it to boot but didn't message back what they did. I am thinking they just got it rebooted by boot menu, selecting the previous kernel. I went in to the grub.conf file and the entries are still there with a default entry pointing to one of the updated kernel files. The entry had it pointing the default to 1 (second entry) which I corrected to the previous working kernel. Entry 1 is the updated kernel file without the smp. I have a P4 and the current running (working) kernel is the ELsmp. 2.6.9-42.0.3.ELsmp I understand that smp is for the hyperthreading and dual processor. When WHM reports server information it shows two processors which I guess is the hyperthreading being detected. Would the server trying to boot with the non smp kernel cause it to fail boot with CPanel? Is it possible they changed something else to make the server boot to the previous kernel even though grub was pointing to the non working kernel file? The working entry in grub is 2.
Thanks

 

[eagle]

This is guesswork, since we have no clue what happened. There should be someone present to watch the boot process.

From my own experience, servers with 'smp' cpu's boot normally on a 'non-smp' kernel.

 

[blkjck]

Yeah but that is more than what anyone at the DC will do. I've been going over the logs and it seems to lock up after starting pure-ftp, the logs stop then restart. If I understand the log, looks like they hooked up a usb keyboard and rebooted it using boot menu. I was comparing the differences in the startup entries. And this below is the main difference at startup. Now that I know about the --once option, I will give the other smp kernel a try. Need to check out the config files for differences. If anyone knows what this error message implies, would appreciate the input. Thanks.

2.6.9-55.0.12.EL
Jan 25 00:36:17 srv01 kernel: found SMP MP-table at 000f4ba0
Jan 25 00:36:17 srv01 kernel: Using x86 segment limits to approximate NX protection
Jan 25 00:36:17 srv01 irqbalance: Can't balance irqs on a uniprocessor system: failed
Jan 25 00:36:17 srv01 kernel: zapping low mappings.


2.6.9-42.0.3.ELsmp
Jan 25 01:38:43 srv01 kernel: found SMP MP-table at 000f4ba0
Jan 25 01:38:43 srv01 kernel: NX (Execute Disable) protection: active
Jan 25 01:38:43 srv01 kernel: DMI 2.4 present.
Jan 25 01:38:43 srv01 kernel: Using APIC driver default
Jan 25 01:38:43 srv01 irqbalance: irqbalance startup succeeded