Hello,

I need to automate installation of CentOS 7.0 using PXE and kickstart. My server is diskless, I’m installing CentOS on a 25GB LUN on a FC SAN
with 4 paths. This LUN is the only one accessible by the server.

Anaconda processes successfully the installation. But on reboot, I get stuck on a prompt after BIOS. I think I have an issue with storage configuration / bootloader configuration…

I can install it using the ISO, with basic configuration. The
/root/anaconda-ks.cfg looks like the following (I striped some part, I
cannot copy past because I have only console access through a secured VPN :-/ ):
firstboot –enable ignoredisk –only-use=mpatha,sdc,sdd,sdb,sda bootloader –location=mbr –boot-drive=mpatha autopart –type=lvm clearpart –all –initlabel –drives=mpatha
%packages
@base
@core
%end

My ks.cfg template is identical (I also cannot attach it unfortunately…).

I tried several things like:
– put sda as boot-drive for bootloader (according to http://osdir.com/ml/general/2014-09/msg52181.html ), but anaconda fails with error that this drive cannot be found. It asks to choose another disk, mpatha is the only available.
– same with dm-0 (as /dev/mapper/mpatha is a symlink to /dev/dm-0) but anaconda also fails
– add to bootloader « –append=”rd.lvm.lv=CentOS/root rd.lvm.lv=CentOS/swap” »

I also found some post on the web saying that in the « ignoredisk »
mpatha mustn’t be used but disk/by-id/scsi-XXXX (chapter 23.3.2 of RHEL
7 installation guide), but I don’t know how to manage that.

Is someone ever did that kind of PXE installation, with boot on san &
LVM, and help me in the kickstart configuration?

Any help would be highly appreciated!

Best regards, Damien