25.3. Adding a Network Device
Network device driver modules are loaded automatically by udev.
You can add a network interface on IBM System z dynamically or persistently.
- Dynamically
- Load the device driver
- Remove the network devices from the list of ignored devices.
- Create the group device.
- Configure the device.
- Set the device online.
- Persistently
- Create a configuration script.
- Activate the interface.
The following sections provide basic information for each task of each IBM System z network device driver. Section 25.3.1, “Adding a qeth Device” describes how to add a qeth device to an existing instance of Red Hat Enterprise Linux. Section 25.3.2, “Adding an LCS Device” describes how to add an lcs device to an existing instance of Red Hat Enterprise Linux. Section 25.3.3, “Mapping Subchannels and Network Device Names” describes how persistent network device names work. Section 25.3.4, “Configuring a System z Network Device for Network Root File System” describes how to configure a network device to use with a root file system that is only accessible through the network.
25.3.1. Adding a qeth Device
The qeth network device driver supports System z OSA-Express features in QDIO mode, HiperSockets, z/VM guest LAN, and z/VM VSWITCH.
Based on the type of interface being added, the qeth device driver assigns one of the base interface names:
- hsin for HiperSockets devices
- ethn for Ethernet features
The value n is an integer that uniquely identifies the device. n is
0
for the first device of that type, 1
for the second, and so on.
25.3.1.1. Dynamically Adding a qeth Device
To add a qeth device dynamically, follow these steps:
- Determine whether the qeth device driver modules are loaded. The following example shows loaded qeth modules:
# lsmod | grep qeth qeth_l3 127056 9 qeth_l2 73008 3 ipv6 492872 155ip6t_REJECT,nf_conntrack_ipv6,qeth_l3 qeth 115808 2 qeth_l3,qeth_l2 qdio 68240 1 qeth ccwgroup 12112 2 qeth
If the output of thelsmod
command shows that the qeth modules are not loaded, run themodprobe
command to load them:# modprobe qeth
- Use the
cio_ignore
command to remove the network channels from the list of ignored devices and make them visible to Linux:# cio_ignore -r read_device_bus_id,write_device_bus_id,data_device_bus_id
Replace read_device_bus_id,write_device_bus_id,data_device_bus_id with the three device bus IDs representing a network device. For example, if the read_device_bus_id is0.0.f500
, the write_device_bus_id is0.0.f501
, and the data_device_bus_id is0.0.f502
:# cio_ignore -r 0.0.f500,0.0.f501,0.0.f502
- Use the znetconf command to sense and list candidate configurations for network devices:
# znetconf -u Scanning for network devices... Device IDs Type Card Type CHPID Drv. ------------------------------------------------------------ 0.0.f500,0.0.f501,0.0.f502 1731/01 OSA (QDIO) 00 qeth 0.0.f503,0.0.f504,0.0.f505 1731/01 OSA (QDIO) 01 qeth 0.0.0400,0.0.0401,0.0.0402 1731/05 HiperSockets 02 qeth
- Select the configuration you want to work with and use znetconf to apply the configuration and to bring the configured group device online as network device.
# znetconf -a f500 Scanning for network devices... Successfully configured device 0.0.f500 (eth1)
- Optionally, you can also pass arguments that are configured on the group device before it is set online:
# znetconf -a f500 -o portname=myname Scanning for network devices... Successfully configured device 0.0.f500 (eth1)
Now you can continue to configure the networketh1
interface.
Alternatively, you can use sysfs attributes to set the device online as follows:
- Create a qeth group device:
# echo read_device_bus_id,write_device_bus_id,data_device_bus_id > /sys/bus/ccwgroup/drivers/qeth/group
For example:# echo 0.0.f500,0.0.f501,0.0.f502 > /sys/bus/ccwgroup/drivers/qeth/group
- Next, verify that the qeth group device was created properly by looking for the read channel:
# ls /sys/bus/ccwgroup/drivers/qeth/0.0.f500
You may optionally set additional parameters and features, depending on the way you are setting up your system and the features you require, such as:portno
layer2
portname
For information on additional parameters, refer to the chapter on the qeth device driver in Linux on System z Device Drivers, Features, and Commands on Red Hat Enterprise Linux 6. - Bring the device online by writing 1 to the online sysfs attribute:
# echo 1 > /sys/bus/ccwgroup/drivers/qeth/0.0.f500/online
- Then verify the state of the device:
# cat /sys/bus/ccwgroup/drivers/qeth/0.0.f500/online
1
A return value of1
indicates that the device is online, while a return value0
indicates that the device is offline. - Find the interface name that was assigned to the device:
# cat /sys/bus/ccwgroup/drivers/qeth/0.0.f500/if_name
eth1
Now you can continue to configure the networketh1
interface.The following command from the s390utils package shows the most important settings of your qeth device:# lsqeth eth1 Device name : eth1 --------------------------------------------- card_type : OSD_1000 cdev0 : 0.0.f500 cdev1 : 0.0.f501 cdev2 : 0.0.f502 chpid : 76 online : 1 portname : OSAPORT portno : 0 state : UP (LAN ONLINE) priority_queueing : always queue 0 buffer_count : 16 layer2 : 1 isolation : none
25.3.1.2. Dynamically Removing a qeth Device
To remove a qeth device, use the znetconf tool. For example:
- Use the
znetconf
command to show you all configured network devices:znetconf -c Device IDs Type Card Type CHPID Drv. Name State -------------------------------------------------------------------------------- 0.0.8036,0.0.8037,0.0.8038 1731/05 HiperSockets FB qeth hsi1 online 0.0.f5f0,0.0.f5f1,0.0.f5f2 1731/01 OSD_1000 76 qeth eth0 online 0.0.f500,0.0.f501,0.0.f502 1731/01 GuestLAN QDIO 00 qeth eth1 online
- Select the network device to be removed and trigger
znetconf
to set the device offline and ungroup the ccw group device.# znetconf -r f500 Remove network device 0.0.f500 (0.0.f500,0.0.f501,0.0.f502)? Warning: this may affect network connectivity! Do you want to continue (y/n)?y Successfully removed device 0.0.f500 (eth1)
- Verify the success of the removal:
znetconf -c Device IDs Type Card Type CHPID Drv. Name State -------------------------------------------------------------------------------- 0.0.8036,0.0.8037,0.0.8038 1731/05 HiperSockets FB qeth hsi1 online 0.0.f5f0,0.0.f5f1,0.0.f5f2 1731/01 OSD_1000 76 qeth eth0 online
25.3.1.3. Persistently Adding a qeth Device
To make your new qeth device persistent you need to create the configuration file for your new interface. The network interface configuration files are placed in
/etc/sysconfig/network-scripts/
.
The network configuration files use the naming convention
ifcfg-device
, where device is the value found in the if_name
file in the qeth group device that was created earlier. In this example it is eth1
. cio_ignore
is handled transparently for persistent device configurations and you do not need to free devices from the ignore list manually.
If a configuration file for another device of the same type already exists, the simplest solution is to copy it to the new name.
# cd /etc/sysconfig/network-scripts # cp ifcfg-eth0 ifcfg-eth1
If you do not have a similar device defined you must create one. Use this example of
ifcfg-eth0
as a template:
/etc/sysconfig/network-scripts/ifcfg-eth0
# IBM QETH DEVICE=eth0 BOOTPROTO=static IPADDR=10.12.20.136 NETMASK=255.255.255.0 ONBOOT=yes NETTYPE=qeth SUBCHANNELS=0.0.09a0,0.0.09a1,0.0.09a2 PORTNAME=OSAPORT OPTIONS='layer2=1 portno=0' MACADDR=02:00:00:23:65:1a TYPE=Ethernet
Edit the new ifcfg-eth1 file as follows:
- Modify the
DEVICE
statement to reflect the contents of theif_name
file from your ccwgroup. - Modify the
IPADDR
statement to reflect the IP address of your new interface. - Modify the
NETMASK
statement as needed. - If the new interface is to be activated at boot time, then make sure
ONBOOT
is set toyes
. - Make sure the
SUBCHANNELS
statement matches the hardware addresses for your qeth device. - Modify the
PORTNAME
statement or leave it out if it is not necessary in your environment. - You may add any valid sysfs attribute and its value to the
OPTIONS
parameter. The Red Hat Enterprise Linux installer currently uses this to configure the layer mode (layer2
) and the relative port number (portno
) of qeth devices.The qeth device driver default for OSA devices is now layer 2 mode. To continue using old ifcfg definitions that rely on the previous default of layer 3 mode, addlayer2=0
to theOPTION
S parameter.
/etc/sysconfig/network-scripts/ifcfg-eth1
# IBM QETH DEVICE=eth1 BOOTPROTO=static IPADDR=192.168.70.87 NETMASK=255.255.255.0 ONBOOT=yes NETTYPE=qeth SUBCHANNELS=0.0.0600,0.0.0601,0.0.0602 PORTNAME=OSAPORT OPTIONS='layer2=1 portno=0' MACADDR=02:00:00:b3:84:ef TYPE=Ethernet
Changes to an
ifcfg
file only become effective after rebooting the system or after the dynamic addition of new network device channels by changing the system's I/O configuration (for example, attaching under z/VM). Alternatively, you can trigger the activation of a ifcfg
file for network channels which were previously not active yet, by executing the following commands:
- Use the
cio_ignore
command to remove the network channels from the list of ignored devices and make them visible to Linux:# cio_ignore -r read_device_bus_id,write_device_bus_id,data_device_bus_id
Replace read_device_bus_id,write_device_bus_id,data_device_bus_id with the three device bus IDs representing a network device. For example, if the read_device_bus_id is0.0.0600
, the write_device_bus_id is0.0.0601
, and the data_device_bus_id is0.0.0602
:# cio_ignore -r 0.0.0600,0.0.0601,0.0.0602
- To trigger the uevent that activates the change, issue:
echo add > /sys/bus/ccw/devices/read-channel/uevent
For example:echo add > /sys/bus/ccw/devices/0.0.0600/uevent
- Check the status of the network device:
# lsqeth
- Now start the new interface:
# ifup eth1
- Check the status of the interface:
# ifconfig eth1 eth1 Link encap:Ethernet HWaddr 02:00:00:00:00:01 inet addr:192.168.70.87 Bcast:192.168.70.255 Mask:255.255.255.0 inet6 addr: fe80::ff:fe00:1/64 Scope:Link UP BROADCAST RUNNING NOARP MULTICAST MTU:1492 Metric:1 RX packets:23 errors:0 dropped:0 overruns:0 frame:0 TX packets:3 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:644 (644.0 b) TX bytes:264 (264.0 b)
- Check the routing for the new interface:
# route Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 192.168.70.0 * 255.255.255.0 U 0 0 0 eth1 10.1.20.0 * 255.255.255.0 U 0 0 0 eth0 default 10.1.20.1 0.0.0.0 UG 0 0 0 eth0
- Verify your changes by using the ping command to ping the gateway or another host on the subnet of the new device:
# ping -c 1 192.168.70.8 PING 192.168.70.8 (192.168.70.8) 56(84) bytes of data. 64 bytes from 192.168.70.8: icmp_seq=0 ttl=63 time=8.07 ms
- If the default route information has changed, you must also update
/etc/sysconfig/network
accordingly.