The new device naming convention aka eth0 is now p2p1

You may or may not have run into this yet. I’ve only seen it a few times and while I understand the reasoning behind it, it plays hell with kickstart (since you have to know the interface name and assume it’s eth0). The behavior can be disabled, apparently through a boot parameter:

biosdevname=0

I’ll test it when I run across a machine that does this to see if it can be handled gracefully in kickstart.

Advertisements

Kickstart without DHCP

Occasionally, you have to kickstart a host and you don’t have a DHCP server available on the network. To setup the network far enough to load a kickstart remotely (via http) change the boot configuration to:

linux text ip=192.168.1.127 netmask=255.255.255.0 gateway=192.168.1.1 dns=192.168.1.10 ks=http://192.168.1.5/kickstart/template.cfg

Alter the above to suit your environment, keeping in mind that the network specified will only be used long enough to retrieve the kickstart. Your template can redefine the values.

DHCP server configuration

This is a bare minimum guide to setup a DHCP server.

1. Install the dhcp package:

yum install dhcp

2. Replace the /etc/dhcpd.conf with the following:

ddns-update-style interim;
ignore client-updates;

subnet 192.168.1.0 netmask 255.255.255.0 {
authoritative;

option routers 192.168.1.1;
option subnet-mask 255.255.255.0;

option domain-name "domain.com";
option domain-name-servers 192.168.1.5;

option time-offset -18000; # Eastern Standard Time

range dynamic-bootp 192.168.1.32 192.168.1.47;
default-lease-time 21600;
max-lease-time 43200;
}

3. Alter the values as needed.
4. Set the server to start on boot:

/sbin/chkconfig dhcpd on

5. Start the server:

/sbin/service dhcpd start

6. Review /var/log/messages for any errors.

Make sure you have port 67/udp open if you are using a host-based firewall.

SSH Multiplexing

Create a new file in ~/.ssh called config and insert the following:

Host *
ControlMaster auto
ControlPath ~/.ssh/master-%r@%h:%p

Reuse an existing SSH connection when you connect to the same host multiple times. Be sure to chmod 600 the config file.

Update: This is also nice to add (not supported on all versions of ssh client), ControlPersist 4h

NFS server and client configuration

If you prefer to do minimal installs (me too) you may have noticed that nfs functions are no longer in the base. Here’s the quick steps to configure a server and clients.

1. On the “server” edit the file /etc/exports (alter to your needs)

/export/media 192.168.1.0/24(rw,async,no_root_squash)

2. Run /usr/sbin/exportfs -a
3. Run /usr/sbin/exportfs

/export/media 192.168.1.0/24

4. Run yum install nfs-utils
5. Run /etc/init.d/rpcbind start
6. Run /etc/init.d/nfs start

To set these to run at boot:
1. Navigate to /etc/init.d
2. Run /sbin/chkconfig rpcbind on
3. Run /sbin/chkconfig nfs on

Now that the server is up, let’s get the clients configured.
1. Run yum install nfs-utils
2. Modify /etc/fstab (alter to your needs)

192.168.1.254:/export/media /usr/local/media nfs defaults 0 0

3. Run /etc/init.d/rpcbind start
4. Run /etc/init.d/netfs start

To set these to run at boot:
1. Navigate to /etc/init.d
2. Run /sbin/chkconfig rpcbind on
3. Run /sbin/chkconfig netfs on

This should result in a mount on the client. Test this by creating files and making sure they appear in the mount point. Obviously this is just a quick and dirty guide and for further details, consult the appropriate documentation.

If you get the weird UID/GID issue (4294967294) then modify the /etc/nfsmount.conf on the client machines. This value is 4 by default and it’s commented, uncomment it, change it to 3 and save it.

Defaultvers=3

Restart the netfs on the client (which will remount as an NFSv3 client and the UID/GID should revert back to 0 (root).