Where does my rpm goes when I update the kernel

hexadecimal_0xF · 12-29-2006, 10:51 PM

Hi!
I have FC5 installed into my laptop

Intel Centrino Dua
2GB RAM

Everything is working just great.

I had installed some rpms to support NTFS file systems and some other.
Now I just updated the kernel using yum, I can see for this new kernel I will have to install all those rpms (why does this happen).

I have some questions to ask:

How can I completely remove my older kernel, which is taking
unecessary space?

Does the rpm get themselves attached to kernel itself. I am
askingbecause I would have to remove the older rpms too (which
I have installed with previous kernel).

I am confused what is architechture of my machine, whether it is
32 bit or 64 bit. Because before automatically updating the
kernel it was 2.6.x.xx-fc5smp but now it is minus smp.

(optional ques) I am running low on disk space, what measure should I take to contain the situation. Do we have anything like add/remove program in FC5.

I would glad to provide more information about my system if you need.

Thank you very much in advance.

PatrickNew · 12-29-2006, 11:33 PM

If I remember correctly, yum is configured to only keep 3 kernels, as a safety precaution. So, you'll never have more than three, but until you hit three yum won't automatically remove them for you. You can manually remove that kernel by using the command "rpm -e kernel-<version you're removing>". As root, of course. The rpm's for ntfs etc. should be updatable with yum.

BTW, wikipedia says that a core duo is strictly 32-bit. As for why the SMP dissappeared, I'm not sure, but you should look into getting a SMP kernel to take advantage of your dual core processor.

jschiwal · 12-30-2006, 12:14 AM

Type in "uname -a" to print out your architecture and kernel information. If you used the distro's software installation program to install your old kernel, you can uninstall it by uninstalling the rpm package it came from.

You can allow use the "-e" option to remove packages. You can search for installed packages like this: rpm -qa | grep kernel.

The rpm's you installed to support NTFS, probably included some kernel modules that were installed to /lib/modules/<kernel version>/. When I update my kernel, I need to reinstall ndiswrapper and my nvidia driver as well.

hexadecimal_0xF · 12-30-2006, 01:12 AM

Thank you very much to both of you for a prompt reply.

Well!! I got the following

Code:

[root@localhost ~]# uname -i
i386
[root@localhost ~]# uname -p
i686
[root@localhost ~]# uname -a
Linux localhost.localdomain 2.6.18-1.2257.fc5 #1 Fri Dec 15 16:06:24 EST 2006 i686 i686 i386 GNU/Linux
[root@localhost ~]#

Please suggest me to use the most suited kernel.

jschiwal · 12-31-2006, 02:21 AM

You have never given your computer a hostname?

Look in /lib/modules/<old-kernel-version>/
This will contain modules for your old kernel.
You can find what package supplied the module with rpm -qf <full/path/name>.
Uninstalling the old kernel itself may also uninstall the packages that supplied the old modules. Be careful however. Using your GUI software management program may be a good idea. It should tell you which packages where dependant on that kernel and give you the option of removing them as well, or aborting.

Also, if you are certain that you will never go back to that kernel version, you could delete the old kernel file and initrd-<version> file from /boot manually. Also edit your grub's menu.lst file to remove the entry as well. You should also be able to remove the source and module directory heirarchy trees as well manually.

However, for a minor version change, make sure there is a later kernel source tree in /usr/src/ and make sure that /usr/src/linux is linked to it.

Just A) go slow and B) be careful; and you should be OK.

PhilTR · 01-24-2007, 07:17 PM

I want to uninstall a newer kernel and go back to an older version so I can install 'NVIDIA-Linux-x86-1.0-9631-pkg1.run'.

'rpm -qa shows:

[root@localhost philtr]# rpm -qa | grep kernel
kernel-devel-2.6.18-1.2798.fc6
kernel-2.6.18-1.2798.fc6
kernel-2.6.19-1.2895.fc6
kernel-devel-2.6.19-1.2895.fc6
kernel-headers-2.6.19-1.2895.fc6

A post on NVNEWS by "zander" (NVIDIA Corp) suggests:

"If you wish to install the NVIDIA Linux graphics driver on a Fedora Core 5/6 system, please ensure that your system meets the following requirements:

* the latest update kernel is installed and in use
* a kernel-devel RPM is installed that corresponds to the kernel that is in use
* the pkgconfig and xorg-x11-server-sdk RPMs are installed
* Xen kernels are not currently supported
* the SELinux environment allows the NVIDIA Linux graphics driver to work (only for drivers older than 1.0-9631 and 1.0-7184 (legacy), respectively)

The first three items above can be addressed with the following commands (they need to be run as root):

# yum install kernel-devel xorg-x11-server-sdk pkgconfig
# yum update
# reboot

If you are using an SMP (multi-CPU and/or multi-core) enabled system in FC5(only), please replace kernel-devel with kernel-smp-devel in the command above."

Now what really has me puzzled is:

[root@localhost philtr]# uname -a
Linux localhost.localdomain 2.6.19-1.2895.fc6 #1 SMP Wed Jan 10 18:32:37 EST 2007 i686 i686 i386 GNU/Linux

Plain and simple, I do not have (or I don't think I have) a multi core CPU. It's a P-IV 2.86 Ghz. When I ran "Software Updater', the newer kernel was one of the pkgs to be updated/installed. Call me dummy.

Nevertheless, running 'yum install kernel-devel xorg-x11-server-sdk pkgconfig' (with the 2.6.18 kernel loaded) shows:

"[root@localhost philtr]# yum install kernel-devel xorg-x11-server-sdk pkgconfig
Loading "installonlyn" plugin
Setting up Install Process
Setting up repositories
livna 100% |=========================| 1.1 kB 00:00
core 100% |=========================| 1.1 kB 00:00
updates 100% |=========================| 1.2 kB 00:00
extras 100% |=========================| 1.1 kB 00:00
Reading repository metadata in from local files
Parsing package install arguments
Resolving Dependencies
--> Populating transaction set with selected packages. Please wait.
---> Package kernel-devel.i686 0:2.6.18-1.2869.fc6 set to be installed
---> Package xorg-x11-server-sdk.i386 0:1.1.1-47.4.fc6 set to be updated
--> Running transaction check
--> Populating transaction set with selected packages. Please wait.
---> Package kernel-devel.i686 0:2.6.18-1.2798.fc6 set to be erased
--> Running transaction check

Dependencies Resolved

=============================================================================
Package Arch Version Repository Size
=============================================================================
Installing:
kernel-devel i686 2.6.18-1.2869.fc6 updates 4.6 M
xorg-x11-server-sdk i386 1.1.1-47.4.fc6 updates 363 k
Removing:
kernel-devel i686 2.6.18-1.2798.fc6 installed 14 M

Transaction Summary
=============================================================================
Install 2 Package(s)
Update 0 Package(s)
Remove 1 Package(s)

Total download size: 5.0 M
Is this ok [y/N]: y
Downloading Packages:
Running Transaction Test
Finished Transaction Test

Transaction Check Error: package kernel-devel-2.6.19-1.2895.fc6 (which is newer than kernel-devel-2.6.18-1.2869.fc6) is already installed"

This seems to suggest that the 2.6.19 kernel is interfearing with the driver install ( I boot the 2.6.18 kernel when trying to install the nVidia drivers).

My question is can I uninstall all the 2.6.19 pkgs w/o interfearing with the 2.6.18 pkgs? Specifically does the 2.6.18 kernel need a 'headers' pkg? I'm not sure if the 2.6.18 kernel ever had an associated header pkg. phil

v00d00101 · 01-24-2007, 09:47 PM

Also to free up space, try purging all the rpm files that tend to hide in /var/cache/yum . There may be a couple located there that can be removed to free up space.

On the nvidia problem have you tried installing it via yum? You need to add the livna first by running the following from a term as root or su:

Code:

rpm -ivh http://rpm.livna.org/livna-release-5.rpm

Now run.

Code:

yum install kmod-nvidia

As for uninstalling the latest kernel, yes you can, but yum will probably try to update it next time you run it. You need to first reboot to an older kernel, then do the rpm -e on the newer one.

PhilTR · 01-25-2007, 07:03 AM

I have livna installed already. Tried to install the nVidia driver via livna (while booted up in the 2.6.18 kernel) but, X server wouldn't load. I'll see if I can post any informative logs. phil

santoshv2005 · 02-01-2007, 12:47 AM

Check the current kernel version with uname -r
check the current kernel version,processor and all details with uname -a.
to check all kernels installed, give the coomand:
rpm -qa | grep kernel
Now check /etc/grub.conf (bootloader configuration file, it shows both the installed kernels).
to remove the old kernel use the command
rpm -e <kernel-version> to remove the older kernel.
Remember to give the kernel-version of the older kernel which you intend to remove.
/etc/grub.conf is automatically updated.

This solution is for RedHat Enterprise Linux, kindly use the appropriate equivalent files
for other distributions.
Hope this solves your problem.
good luck.

santoshv2005 · 02-01-2007, 12:50 AM

Dear Friends
kindly give me the labs or tutorials , pdfs for implementing LDAP Server and client configurations in RHEL4.
(RedHat Enterprise Linux version4 )