2

When trying to provision a CentOS 7 system with the RedHat Satellite 6.2.9 I got the following error:

...

ValueError: new value non-existent xfs filesystem is not valid as a default fs type

Pane is dead

Can anyone help?

Community
  • 1
Adail Junior
  • 141
  • 4

2 Answers2

2

Well after a few hours I solved the problem.

Despite the fact that error message talks about non-existent file system, the problem itself doesn't have nothing to do with the file system type. The real problem is that the vmlinuz/initrd.img files that is been delivery to the machine during the first boot process is different from the ones that are delivered to Anaconda when it starts the installation process.

In RedHat Satellite 6.2.9, you have to find what are the correct vmlinuz/initrd files in the pulp repository:

$cd /var/lib/pulp/published/yum/master/yum_distributor/
$for i in `find . -name "*vmlinuz*"`; do md5sum $i; done
e26fb62adf407f112b47d63eff6bd23c  ./Library-Ceph-CentOS-7-CentOS-7-os/1508160648.68/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS-17_0-CentOS-7-CentOS-7-os/1510159832.92/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS-18_0-CentOS-7-CentOS-7-os/1512565858.98/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Producao-Ceph-CentOS-7-CentOS-7-os/1508161486.78/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS-21_0-CentOS-7-CentOS-7-os/1517253969.47/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS-22_0-CentOS-7-CentOS-7-os/1517255117.41/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS-23_0-CentOS-7-CentOS-7-os/1521811840.7/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Producao-Generico_CENTOS-CentOS-7-CentOS-7-os/1527697980.04/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Producao-Generico_CENTOS_POSTGRES-CentOS-7-CentOS-7-os/1521834748.44/images/pxeboot/vmlinuz
e26fb62adf407f112b47d63eff6bd23c  ./Generico_CENTOS_POSTGRES-4_0-CentOS-7-CentOS-7-os/1521833864.47/images/pxeboot/vmlinuz

All of them have the same md5 because they point to the same file. The same need to be done with the initrd.img:

$for i in `find . -name "*initrd.img*"`; do md5sum $i; done

Then replace with the correct files in /var/lib/tftpboot/boot directory:

$cd /var/lib/tftpboot/boot
$ls |grep CentOS 
CentOS-7.2-x86_64-initrd.img
CentOS-7.2-x86_64-vmlinuz
CentOS-7.3-x86_64-initrd.img
CentOS-7.3-x86_64-vmlinuz
CentOS-7.4-x86_64-initrd.img
CentOS-7.4-x86_64-vmlinuz
CentOS-7-x86_64-initrd.img
CentOS-7-x86_64-vmlinuz
CentOS_Linux-7-x86_64-initrd.img
CentOS_Linux-7-x86_64-vmlinuz
$sudo cp /var/lib/pulp/published/yum/master/yum_distributor/Generico_CENTOS-17_0-CentOS-7-CentOS-7-os/1510159832.92/images/pxeboot/vmlinuz CentOS-7.2-x86_64-vmlinuz CentOS-7.2-x86_64-vmlinuz

Do it for every vmlinuz and initrd.img file that relates with the OS system with problem.

Adail Junior
  • 141
  • 4
0

This happens with all Satellite versions (and Foreman) when kernel/initramdisk changes after new OS release and/or after repo sync or is downloaded incorrectly. We improved this issue a bit in Foreman 1.20 (Satellite 6.5) but it's still not solved. A preliminary solution has been delivered in Satellite 6.6 (beta) and Foreman 1.22: https://projects.theforeman.org/issues/26709

WORKAROUND: Delete initramdisk/vmlinuz on TFTP server and re-enter build mode to download these.

lzap
  • 2,032
  • 2
  • 14
  • 13