Hi again.
I'm pleased to say that my experimentation with ocfs2 has been very successful and I'm considering doing some real work with it.
My current question is related to virt-manager and the use of kickstart to install paravirtualised domU's. I'm happily doing fully-virtualised installs with this tool but due to poor network performance I thought it time to do some para-virtualised testing. In the dialog where I'm asked to supply the URL for the install repository and the kickstart location I'm unable to use the "file://..." construct so I'm left with a selection that only includes locations on the network.
Now my network interface information for the domU is in the kickstart file so when I move forward in the tool there's a message saying that its trying to determine a network address, presumably via DHCP (I am unable to put a dhcp server on this network). From the command line I can use wget against the kickstart file at the URL I'm providing so that's OK.
I'm obviously overlooking something but its currently got me puzzled. So how is this supposed to work? Am I supposed to do the network configuration for the installation somewhere else?
Brett
On Wed, Sep 10, 2008 at 08:10:27PM +1000, Brett Worth wrote:
Now my network interface information for the domU is in the kickstart file so when I move forward in the tool there's a message saying that its trying to determine a network address, presumably via DHCP (I am unable to put a dhcp server on this network). From the command line I can use wget against the kickstart file at the URL I'm providing so that's OK.
I'm obviously overlooking something but its currently got me puzzled. So how is this supposed to work? Am I supposed to do the network configuration for the installation somewhere else?
The network information is passed with -x "boot parameters" just as if you were booting a CD but wanted to do an unattended network install.
I'm using a templatized script, but the resulting cmdline version is the following:
virt-install --nonsparse -b xenbr1.246 --vncport=5905 -p \ -l http://ftp.sunet.se/pub/Linux/distributions/centos/5/os/i386 \ -f /var/lib/xen/images/host05.xvda -s 20G -n host05 -r 256 --vnc \ -x "cmdline ks=http://repo-server/kickstart/host05.cfg ksdevice=eth0 \ ip=X.X.X.X \ netmask=X.X.X.X \ gateway=X.X.X.X \ dns=X.X.X.X
-- Fridh
Mikael Fridh wrote:
The network information is passed with -x "boot parameters" just as if you were booting a CD but wanted to do an unattended network install.
I'm using a templatized script, but the resulting cmdline version is the following:
virt-install --nonsparse -b xenbr1.246 --vncport=5905 -p \ -l http://ftp.sunet.se/pub/Linux/distributions/centos/5/os/i386 \ -f /var/lib/xen/images/host05.xvda -s 20G -n host05 -r 256 --vnc \ -x "cmdline ks=http://repo-server/kickstart/host05.cfg ksdevice=eth0 \ ip=X.X.X.X \ netmask=X.X.X.X \ gateway=X.X.X.X \ dns=X.X.X.X
Thanks Mikael. I'm now off and running...
Brett