PXEboot files from 9.1 fail

Hello,

I’ve tried to setup a PXE boot server using the pxeboot files provided here :
https://download.rockylinux.org/pub/rocky/9/BaseOS/x86_64/kickstart/images/pxeboot

It boots fine, but I have to wait for a job on /dev/zram0 to fail during 1m30s :

And then, when anaconda starts, the storage module fails to start :

It seems like this is due to a missing XFS module in initrd :

I can’t find any bug report about that, so either I should fill one or I’m doing something wrong :thinking:

Thanks for your help !

The XFS message you’re looking at is not the issue, it’s a symptom.

With that being said, my PXE environment is not having any issues. This leads me to believe that the boot arguments you’re using are not valid and/or your environment is not sufficiently set up.

I would check the following:

  • Your VM has enough RAM (the minimum as stated by Red Hat is 4gb, you may get away with 3gb, but this is up to you to determine this)
  • Your boot arguments are correct (I recommend that inst.repo and inst.stage2 both be specified)







Hello Louis,

Thank you very much for your reply AND your amazing work on Rocky Linux !
You nailed it. I was missing the inst.stage2= boot option.
By chance, it used to work with 9.0 vmlinuz/initrd files without specifying that runtime image.

I owe you one !

Cheers.

1 Like

Good to hear! I might be misremembering, but there was a point that inst.repo used to work just fine and the stage2 would be automatically determined… I don’t know if this was changed behavior, but I’ve always just used both because I’ve been bitten in the past.