That’s good to hear! We tried some different search techniques, as apposed to the usual way of simply entering words into Google. You might want to check the follwing pages out (or anyone else for that matter):

Basic Sarching
Advanced Searching

Glad to hear you had success after all. Perhaps you would like to share the additional packages you had to install, if there were any?

I used to have the xen kernel, I have removed and gone to a standard kernel and then recreated the symbolic link to match.

I still get the “What is the location of the directory of C header files that match your running kernel” error. I have run update and rebooted several times. Any thoughts?

Welcome to the world of linux! Your question makes perfect sense! My article instructs you to create that symbolic link, to quickly get everything up and running, without a reboot.

If you started up your machine, installed VMware, installed all it’s other required RPMS, installed kernel headers, installed kernel devel files, and then rebooted, before finally running vmware-config.pl, vmware-config should find the location of your header files and build the kernel module automatically.

You probably have but, have you tried rebooting? It would also be a good idea to install the latest kernel, kernel headers and kernel devel. At a console simply type (as root):

yum -y update

If the include directory, in /usr/src/linux is a file, then there is a problem with your symbolic link. If you copied the line of code from my article, remember you need to replace with the directory name of your current running kernel. If you use the Tab key on your keyboard to complete directory names for you, you shouldn't need to know the name. For example:

ln -s /lib/modules/ /build/include include

You press it twice just in case there are multiple directories, as each will become listed. To choose the directory with the highest version number, type in enough characters to help the tab key auto complete the directory name for you, once it does, continue with the rest of the command above.

The directory name "linux" in /usr/src/linux is a phantom name, in that it actually represents the current version of your running kernel (running yum -y update will change this). If it continues to show linux, instead of the current running version of your kernel, when you run vmware-config.pl, then it is possible you don't have everything installed.

Some people will argue against this, but I always install the kernel headers, and the kernel devel files, make sure you have too. ]]> http://www.mokonamodoki.com/vmware-server-on-centos-5#comment-25 cchristm Fri, 02 May 2008 16:01:47 +0000 http://www.mokonamodoki.com/vmware-server-on-centos-5#comment-25 I am having a problem at the location of C header files as well. I have done everything you have said, but the error I am getting is in regards to the symbolic link. I guess the kernel builder is looking for a directory called include, but the include in /usr/src/linux is a file? (Sorry if that doesn't make sense, I've only had linux for a week.) At any rate, I cannot get past this point, any ideas? Thanks! I am having a problem at the location of C header files as well. I have done everything you have said, but the error I am getting is in regards to the symbolic link. I guess the kernel builder is looking for a directory called include, but the include in /usr/src/linux is a file? (Sorry if that doesn’t make sense, I’ve only had linux for a week.) At any rate, I cannot get past this point, any ideas? Thanks!

Thanks for the help!

The fact that you are using a Xen kernel, means things will be different. I was writing in regards to a standard kernel (not x64, Xen or SMP).

Basically, you are using a Xen kernel (hence the 2.6.18-53.1.14.el5xen), try rebooting with a standard kernel (or install one), or ammend my steps for your kernel version.

Regardless of hardware type, you should be able to install a stock kernel.

Having said that, if you want to use Xen, unless you want to run Windows Operating Systems, simply use the Xen Virtual Software instead of Vmware. If you want to use Vmware, using it with a Xen kernel won’t offer you any increase in performance. In fact it will decrease, because the Xen kernel will reserve memory for itself for virtual machines, taking away from the memory that could be used for Vmware Virtual Machines.

Only problem is that it is not working for me. you say that there will be a /build/include directory path under the /lib/modules/ but there is not. I installed the headers and devel packages, but the only place I see an include directory (created after the devel and headers were installed) is in /usr/source/kernals//include

If I use that directory, I’m told that the kernel headers are version 2.6.18-53.1.14.el5 and that the running version is 2.6.18-53.1.14.el5xen. The include directory is in fact under a directory with the “xen” extention, so what’s up?

Centos 5, Vmware server 1

We should try to encourage the use of the site by everyone in the class. Perhaps if everyone knew where the forum was, they would talk there?

After installing the devel and header packages, you should have a directory located at:

/lib/modules/your current running kernel/build/include
“linux” is just a wild card directory name that is replaced by the directory name of your running kernel.

Sometimes the vmware scripts detect it automatically, other times they don’t, thus we create a symbolic link to the correct location, so that it can find it in it’s “default” location (/usr/src/linux/include)

I updated the guide today, so it might be worth giving it another shot, if you are still having troubles.

Thanks for you comment! What you have said is completely valid and true, and other readers should note Darkness’s comment.

I did state my reasons for installing the kernel headers. I’m sure you have had your moments where you wished you had the kernel headers. More on that below.

Perhaps the installation of libX11-devel could be left out, but it’s not hurting the installation either way.

Being the avid user of linux that you certainly seem to be, I am sure you would know that sometimes the default install of a system, is not always the same as the next.

It is assumed that someone who finds their way to this article, is experiencing the problems that I have pointed out.

While this may sound a little strange, you may find it interesting. Before writing this article, I performed a default “server” install of CentOS 5 three times (on the same machine), and three times I was unable to get the vmware kernel module to compile, until I installed the kernel headers. Even after downloading all the updates and rebooting. Installing the kernel development files, and rebooting again.

I’m sure you will agree that linux can be strange sometimes. So for people who may not have as much experience as you and I, why not cover all the bases for their sake, just in case?

Note xinetd was not installed by default, but is needed by VMware server.

Also note that neither libX11-devel nor kernel-headers are necessary in this case. The -devel package for your kernel (i.e., kernel-devel) is definitely necessary.

My vmware-config.pl knew that kernel headers are in /lib/modules/`uname -r`/build/include, I did not need that symbolic link.