RH 7.3 PXE boot initrd-switch-root failed

rainmk · 01-30-2017, 10:58 AM

Using RHEL 7.3

I am using hammer for unattended installations. After adding new host entry in hammer I see

that thru PXE
Kernel loads
mounts ramfs
all devices vga nic etc get initialized

but then it fails to get the KS file. I do not even see an attempt from client.

this is my pxe conf file

timeout 5
default GRUMPY1
label GRUMPY1
kernel /boot/vmlinuz inst.ks=http://mymachine:80/unattended/provision?token=7a430427-039c-44ec-bea1-9773c18508e5 network inst.ks.sendmac biosdevname=1 audit=1
initrd /boot/initrd.img ip=dhcp
IPAPPEND 2

My machine drops to dracut shell throwing following exception.

initrd-switch-root.service - Switch Root
Loaded: loaded (/usr/lib/systemd/system/initrd-switch-root.service; static; vendor preset: enabled)
Active: failed (Result: exit-code) since Thu 2017-01-26 20:02:34 UTC; 26min ago
Process: 638 ExecStart=/usr/bin/systemctl --no-block --force switch-root /sysroot (code=exited, status=1/FAILURE)
Main PID: 638 (code=exited, status=1/FAILURE)

Jan 26 20:02:34 GRUMPY1 systemd[1]: Starting Switch Root...
Jan 26 20:02:34 GRUMPY1 systemctl[638]: Failed to switch root: Specified switch root path /sysroot does not seem to be an OS tree. os-release file is missing.
Jan 26 20:02:34 GRUMPY1 systemd[1]: initrd-switch-root.service: main process exited, code=exited, status=1/FAILURE
Jan 26 20:02:34 GRUMPY1 systemd[1]: Failed to start Switch Root.
Jan 26 20:02:34 GRUMPY1 systemd[1]: Unit initrd-switch-root.service entered failed state.
Jan 26 20:02:34 GRUMPY1 systemd[1]: Triggering OnFailure= dependencies of initrd-switch-root.service.
Jan 26 20:02:34 GRUMPY1 systemd[1]: initrd-switch-root.service failed.

When I curl the token url I am able to get the KS file

# curl http://mymachine:80/unattended/provi...1-9773c18508e5

install
codlin
url --url http://mymachine:80/pulp/repos/JUNK/...6_64/kickstart
lang en_US.UTF-8
keyboard us

network --onboot yes --device ksdevice=bootif --bootproto static --ip 10.1.19.54 --netmask 255.255.255.0 --nameserver 10.1.19.23 --noipv6 --hostname GRUMPY1

authconfig --enableshadow --passalgo=sha512 --kickstart

rootpw --iscrypted <encryptedpassword>

# Reboot after installation
reboot
firewall --service=ssh
selinux --enforcing
timezone --utc Etc/UTC

bootloader --location=partition --driveorder=sda --append="crashkernel=auto rd_NO_MD rd_NO_DM rd_NO_LUKS rd_NO_LVM LANG=en_US.UTF-8 numa=off " --iscrypted --password=<encruptedpassword>

zerombr
clearpart --all --initlabel --drives sda

%include /tmp/diskpart.cfg

I have gone thru these docs

https://access.redhat.com/documentat...art-howto.html

https://access.redhat.com/documentat...ver-setup.html

but no resolution

MensaWater · 01-30-2017, 02:46 PM

You might want to change your title from "RH 7.3" to "RHEL 7.3". The former describes a very old RedHat version and the latter describes the current RedHat Enterprise one. Some folks are still using RH 7.3 (though only God knows why) and lots of history on the web references that so using that name will confuse things more and more as time goes by and might get your posts ignored by people thinking you're using the older version.

John VV · 01-30-2017, 06:06 PM

the very old and very long dead RH7 ( think 1998 ) is well past it's lifetime

now after the name change back in 2003 if you are referring to RHEL7 , that's different

Quote:

for unattended installations

then you DO have the required and paid for server contract license
they range from $799 to over $4000+ per year depending on the servers and CPU's

have you contacted redhat support ?
you are paying for it with the required license

rainmk · 02-01-2017, 10:32 AM

Thanks guys I have figured out the issue. It was UEFI messing with my setup. Now my machine comes up half way and stops at anaconda "raise VlaueError("new value %s is not valid as a default fs type".

And to answer your question no I have not created a ticket with Redhat. Creating a ticket is easy part but getting an approval to create a ticket is tricky.

Jjanel · 02-02-2017, 10:55 PM

Maybe ['WILD guess']: https://access.redhat.com/solutions/697913 (maybe undef fs specified?)
"rhel7: installation fails with "Pane is dead" and "new value non-existent xfs filesystem is not valid as a default fs type"
Solution Verified... [Subscriber exclusive content] ... (that's all I 'get')

Or, better (PXE) maybe this ... The only clue was found in dmesg on the half-booted machine.

rainmk · 02-03-2017, 10:46 AM

Thanks Jjanel; I have already seen that. I do not think that is my issue.

I have taken the vmlinuz and intrd directly from RHEL7.3 DVD. These files were located under /images/pxeboot

-rw-r--r-- 1 jag sky 49M Oct 19 17:41 upgrade.img
-rw-r--r-- 1 jag sky 43M Oct 19 17:40 initrd.img
-rw-r--r-- 1 jag sky 5.2M Oct 19 11:27 vmlinuz
-rw-r--r-- 1 jag sky 664 Oct 19 18:30 TRANS.TBL

Any ideas what ami doing wrong.