Problem: Kickstart hangs indefinitely at checking storage
Expected outcome: Sails through and images a VM with Rocky Linux 10.1
Foreman and Proxy versions: 3.15
Foreman and Proxy plugin versions:
Distribution and version: Rocky Linux 9.6
Other relevant data:
My boss wants me to test out Rocky Linux 10.1 in our environment. I’ve been wracking my brain on this one as anaconda logs haven’t been very helpful. So when trying to deploy it just gets stuck here:
The weird thing is if I assign 10.1 to the host, and associate it with my old hostgroup for Rocky 9 which has Rocky 9 OS, it will install 10.1 sans any of my custom stuff in the kickstart.
If I try to assign it to a new Rocky 10 group that is assigned Rocky 10 OS, it fails at the above.
I can provide more details if anyone asks. Literally the hostgroup is the difference between it stalling in the kickstart or getting through to the end, albeit without any of my custom kickstart snippets, but with Rocky 10.1. Is there anything about support for Redhat 10 in Foreman 3.16 and Katello 4.18? I will happily update my server if that’s the difference.
The Foreman version should not be the issue here. On a base level (possible required adjustments to the kickstart file aside), RHEL10 can be installed with even older Foreman versions.
“Just changing the hostgroup” fixing it can mean a ton of things. Have you double-checked that your hostgroups are setup identically apart from using EL10 instead of EL9 ISO? Have you taken a look at the rendered KS files for both hostgroups and checked for differences there?
Without any error message, stuff like this is hard to analyze remotely, but in my experience, most installation problems come from within the KS file/template, so I would start to see what changes there when you switch the hostgroup.
Alright, thanks for your feedback. We’ve been using this server successfully to manage Rocky 9 for a couple of years. I actually maintain the Kickstarts and snippets through puppet agents and a gitlab project. I just cloned the Rocky 10 template and edited out the one thing that validator complained about which was the “auth” line which set sha512. According to AI that’s handled automatically now. And I had to edit out one sleep for loop for mounting /mnt/sysimage in the pre phase. Other than that it’s identical to the working kickstart that is applied to a different hostgroup and the repo and media URLs are handled with parameters and I”ve confirmed I can browse to them by copying and pasting what I see in the preview. And like I said, with the different hostgroup it’ll actually image Rocky 10.1.
Good to know that my version of foreman isn’t the issue. As I said the anaconda logs aren’t being super helpful. I’ll see if I can increase verbosity and examine them more carefully. It’s a rather frustrating problem to have. I examined a working host group side by side with the new Rocky 10 one and AFAIK there are no differences. I thought there might be a missing parameter or something. Anyway, I’ll get back to it. If I find the solution I’ll post it here.
