Hi
I've been attempting to pxeboot centos 7.7.1908 using a kickstart file to partition disk and install packages, but am having some problems and was wondering if anyone would be able to help.
The machine I am installing the OS on to is a DEL PowerEdge R610.
------------------------------------
What I have done so far:
- Downloaded CentOS-7-x86_64-DVD-1908.iso, mounted the iso and copied the files to use as a template
- Created a directory (/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolinux) with the following files copied from the downloaded iso:
boot.cat
boot.msg
grub.conf
images
efiboot.img
pxeboot
TRANS.TBL
initrd.img
isolinux.bin
isolinux.cfg
ks
test_centos7_ksfile
memtest
Packages
<contains all packages from iso Packages directory>
repodata
03dde1157310b4a6d8e9128de19651c147a71150d406f3efaaf12f7eb83a1f35-primary.xml.gz
177815eb15f48fcab91407f00675f26cbc13f07601c7d971b6255957faa2d1e0-other.sqlite.bz2
17f82b7ec55c8c3c30bd0093d3241f1824aa4dafde0081f9a31829dbb7e3738e-filelists.sqlite.bz2
4af1fba0c1d6175b7e3c862b4bddfef93fffb84c37f7d5f18cfbff08abc47f8a-c7-x86_64-comps.xml.gz
521f322f05f9802f2438d8bb7d97558c64ff3ff74c03322d77787ade9152d8bb-c7-x86_64-comps.xml
7a2197223a2689b44594b09351dead84fe2d8cd0459be3f885bb985bb28af7ca-primary.sqlite.bz2
84ff0ada5bdbf970afd2791ab1850d0596e41b5442e1ec3f7073974b222b7be9-filelists.xml.gz
a7120e4fc4c651382e22fb4c126798235095a1ce58c2c0783cc6d31c2672f544-other.xml.gz
repomd.xml
repomd.xml.asc
TRANS.TBL
splash.png
TRANS.TBL
vesamenu.c32
vmlinuz
- Specified required server-name and filename in dhcpd.conf of dhcp server (this looks to be working ok)
- Set up tfptserver with syslinux bootloader in the root directory(/var/lib/tftpboot/pxelinux.cfg/default) - (this all works for centos 6 and indeed for a manual centos 7 install, so I assume it is ok)
- copy files from iso in images/pxeboot to /var/lib/tftpboot/pxebootCentOS7_7_1908 (this is then passed as the kernel and initrd image in the pxe config file - see below)
- Put the following boot information in /var/lib/tftpboot/pxelinux.cfg/default:
label Centos7_7_1908_vanilla_kstest
menu label Install CentOS 7_7_1908 x86_64 vanilla ks test
kernel pxebootCentOS7_7_1908/vmlinuz
append initrd=pxebootCentOS7_7_1908/initrd.img ip=dhcp method=nfs:nfsvers=3:10.168.6.206:/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolinux ks=nfs:10.168.6.206:/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolinux/ks/test_centos7_ksfile biosdevname=0 net.ifnames=0 bootdev=link ksdevice=eth0 selinux=0 inst.repo=nfs:nfsvers=3:10.168.6.206:/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolinux
------------------------------------
When I try pxe-booting using the kickstart file the boot load seems to pick up the vmlinuz and initrd images from pxebootCentOS7_7_1908 ok, but the installation fails and ends up at a dracut prompt with the following message:
Could not boot
/dev/root does not exist
Starting Dracut Emergency Shell
Warning /dev/root does not exist
Generating "/run/initramfs/rdsosreport.txt"
Entering Emergency Mode. Exit the shell to continue
Type "journalctl" to view system logs
Just prior to that I see:
61.663700] lb-cam-bca-kstest.cam.broadcom.com dracut-initqueue[838]: mount.nfs: /run/install/repo is busy or already mounted
[ 61.664774] lb-cam-bca-kstest.cam.broadcom.com dracut-initqueue[838]: Warning: Couldn't mount nfs:nfsvers=3:10.168.6.206:/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolin
ux
[ 61.687565] lb-cam-bca-kstest.cam.broadcom.com dracut-initqueue[838]: Warning: no suitable images
then a list of messages like this:
Warning: dracut-initqueue timeout - starting timeout scripts
Even though /var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_testks/isofiles/isolinux was successfully mounted earlier and is still mounted on entering the shell.
Looking at the system logs two sets of messages come up as errors (or are highlighted in red, so I assume are errors):
system-udevd[728] could not find module by name='ipmi_si'
system-udevd[728] could not find module by name='ipmi_devinf'
system-udevd[728] could not find module by name='ipmi_msghandler'
and:
iscsid iSCSI daemon with pid=793 started!
can't open InitiatorName configuration file /etc/iscsi/initiatorname.conf
Warning: InitiatorName file /etc/iscsi/initiatorname.iscsi does not exist or does not con...
can't open InitiatorAlias configuration file /etc/iscsi/initiatorname.iscsi
can't open iscsid.safe_logout configuration file /etc/iscsi/iscsid.conf
I have attached the kickstart file(test_centos7_ksfile.txt), /run/initramfs/rdsosreport.txt, the output of journalctl(journalctl.txt) and the output of dmesg(dmsg.txt)
Has anyone seen this problem before ? or know what might be happening here ?
Given I can install manually from the vanilla iso files maybe something wrong with kickstart or location of files ?
....
Further Notes:
If I do a pxe boot from a copy of the raw iso image using the following in pxe config:
label Centos7_7_1908_vanilla
menu label Install CentOS 7.7.1908 x86_64 vanilla
kernel pxebootCentOS7_7_1908/vmlinuz
append initrd=pxebootCentOS7_7_1908/initrd.img ip=dhcp method=nfs:nfsvers=3:10.168.6.206:/var/www/html/serverinstallation/installationimages/centos/7/7.7.1908/vanilla_fromDVD/isofiles biosdevname=0 net.ifnames=0 ksdevice=eth0 selinux=0
I get to the installation pages and can install manually (at least a minimal install) OK.
The tftboot server I am using is actually centos 6:
uname -a
Linux lb-cam-bca-fs 2.6.32-642.el6.x86_64 #1 SMP Tue May 10 17:27:01 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux
I do not think this should matter in any way, but just in case is that ok ? Or do I have to use a centos 7 server ?
In the logs there are some references to IPv6:
[ 8.569434] localhost kernel: IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[ 11.738915] localhost kernel: IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[ 13.969317] localhost kernel: IPv6: ADDRCONF(NETDEV_UP): eth1: link is not ready
[ 29.370400] localhost kernel: IPv6: ADDRCONF(NETDEV_UP): eth2: link is not ready
[ 44.774693] localhost kernel: IPv6: ADDRCONF(NETDEV_UP): eth3: link is not ready
which I thought I had disabled in the kickstart file using the --noipv6 network option. Is that not the correct way to do it ?
