How can I troubleshoot a failing Linux kickstart installation?

Question

I am having some trouble getting my pxebooting kickstart process working.

The script is known to work when installed via a DVD with the local media on the disk. I have updated the script to work with a remote repository and am leveraging PXEBooting so this can be leveraged for my enterprise. The script executes fine up until it starts to download packages.

Checking the log files on the web server hosting the repository, the first file is downloaded successfully with a 200 HTTP code. But the server running the kickstart indicates a failure and attempts to download the package again. I have confirmed this on the web server, as I see multiple requests for the same package repeated, all with the 200 HTTP code. But kickstart indicates that the download failed.

I am using CentOS 5. I copied the entire first DVD (I do not need the OpenOffice suite from the second) to the location on the web server, so the repodata already exists.

I have been able successfully download software from this repository using other systems that have already been built.

No errors appear in any of the log files that I have found on the kickstarted server, no messages are output to the screen. I have not been able to find any means of debugging this issue.

I am hoping that someone here can provide a link or information on how to attain more detailed debugging information to resolve this.

Thank you in advance.

Solution

Can you post an excerpt of your kickstart? What is the relationship between the repository and the system you're building? Same subnet?

I had a period of kickstart installation problems during the middle of the CentOS 5 series. The best thing to do from your standpoint is to check the other virtual terminals during the installation. Are you running the installation in graphical (X Windows) or text mode?

Here's what the different virtual terminals display. You should be able to debug from there.

Alt-F1
The installation dialog when using text or cmdline

Alt-F2
A shell prompt

Alt-F3
The install log displaying messages from install program

Alt-F4
The system log displaying messages from kernel, etc.

Alt-F5
All other messages

Alt-F7
The installation dialog when using the graphical installer

At some point, I was unable to resolve a kickstart performance issue. I ended up changing the installation method to NFS and the issues disappeared. See: CentOS 5.5 remote kickstart installation stalls at "Starting install process." How to debug?

Comments

[becomingwisest]

Try hitting ALT+F2 and ALT+F3 on the server you are kickstarting to see if it has any additional information that might help.

[user2751502]

If you download the package from that URL, do you get a valid package?

[Nick V]

I have tried ALT+F2 and ALT+F3. F2 brings me to the BusyBox prompt, but no error messages. I have looked through the entire filesystem and nothing. The install logs under /mnt/sysimage/root/ do not even contain information.

[Nick V]

I am able to download a package from the repository without error. The issue only exists when I attempt to use the repository for installation. I even attempted to run a manual installation from the repository. The image/stage2.img file is loaded, but when it comes time to download packages the behavior is the same. At this point, I have loopback-mounted the first DVD but the same issue exists. (using CentOS 5.8)

[Nick V]

For the same mystical reason, the NFS change resolved the issue.