Avoiding Gotchas When Upgrading Kernel Via YUM

Latest response

So I'm in the process of catching up our RHEL servers on patches and in the process, upgrading the kernel.

So far, so good. Except I've got one hangup that I'm working thru and have a case open on it. However I have some questions, and it seems almost too easy to run yum update kernel and yum update kernel-headers.

First question, is how do I know if an existing kernel is using specialized drivers and/or software? If it does, how do I know how to handle the update so it hopefully goes smooth?

Second, after I update the kernel, I'm checking under /boot for the latest vmlinuz, initrd and under /boot/grub/grub.conf to see if an entry has been placed under there. Is there anything else that I should be checking?

Also the previous admin wasn't enabling kdump. Is anyone using this to troubleshoot kernel issues when updating?

thanks

Responses

Hi,

I am new to Linux system administration (been a user of Linux for years, though) but now looking after my own workstation. I am very interested to hear any answers to this post, as I have been unfortuantely impacted with a "yum update" just recently. I have a new system, but it came from DELL with RHEL 7.0. On advice of RHEL folks I did a "yum update" to get to RHEL 7.2. Sounds good. Update was "successful", but then my system would not boot. After various struggles I was advised that RHEL cannot help as it is a kernel issue, and to please talk to DELL. DELL advised "do a factory reinstall, something has gone wrong with what was put on the system since factory install". I had installed virtually nothing (just some application software, nothing into the O/S). Well, I did not know what else to do and had no more advice from either RHEL or DELL, so I reluctantly did the factory reset, lost all of my setup so far, including initial user data on the system (was advised the Raid disks where my user data was stored, separate from boot SSD drive, would not be touched by the factor reset.... wrong). My lost user data was not yet backed up... my fault, and I will not make this mistake again.

I now learn from my own reading that there was an easy way to undo my yum update (rather than do a factor reset to 7.0), which would have saved me a lot of grief. Anyways, after factory reset on squeaky clean 7.0 system, I redid the "yum update" and watched all the details . During the update there is a list of complaints about nvidia and kernel (I missed this list as it raced by on my first update attempt). I believe this is the cause of the fail to boot after my first attempt to upgrade to 7.2. My Nvidia graphics card (Quadro K2200) seems is supported by 7.0 + whatever custom driver installation DELL included in factor distribution, but this driver does not work as-is with out-of-the-box RHEL 7.2 kernel.

So long and short of it, yum update is not so foolproof. I was lead to believe it would update no problems, but no one thought to ask about my graphics driver (and I did not know enough to worry about it). My fault.

All that said, I still want to get my system to 7.2 rather than the older 7.0. Anyone with advice on support for nvidia quadro + RHEL 7.2 greatly appreciated. I tried to build nvidia driver as they instructed from their website, but it did not compile properly (lots of compile warnings and errors, failed to build message at end of driver install log) with the new 7.2 kernel in place outside of X11 (and I am not sure if I uninstalled the previous driver properly). Is there an rpm or yum support for nvidia + RHEL 7.2?

Any help greatly appreciated, and thanks.

Paul

Hi Paul,

My suggestion for you is to consider using the kmod-nvidia package offered by ELRepo [1]. Take a look at the nvidia-detect package [2]. It will tell you which version is suitable for your nvidia card.

[1] http://elrepo.org [2] http://elrepo.org/tiki/nvidia-detect

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.