After Replacing a System Board on Oracle Big Data Appliance 'ifconfig eth0' Returns "eth0: error fetching interface information: Device not found"

Asset ID:	1-72-1944741.1
Update Date:	2017-07-04
Keywords:

Solution Type Problem Resolution Sure

Solution 1944741.1 : After Replacing a System Board on Oracle Big Data Appliance 'ifconfig eth0' Returns "eth0: error fetching interface information: Device not found"

Applies to:

Big Data Appliance Hardware - Version All Versions and later
Linux x86-64

Symptoms

After replacing the system board on a BDA server eth0 disappears.

On the node where the system board was replaced, 'ifconfig eth0' returns: eth0: error fetching interface information: Device not found

# ifconfig eth0

eth0: error fetching interface information: Device not found

On all other nodes, 'ifconfig eth0' returns output as expected, like:

# ifconfig eth0

eth0   Link encap:Ethernet HWaddr 00:21:28:E7:83:4E
         inet addr:*.*.*.* Bcast:*.*.*.255 Mask:255.255.224.0
         UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
         RX packets:7148256 errors:0 dropped:0 overruns:0 frame:0
         TX packets:4236 errors:0 dropped:0 overruns:0 carrier:0
         collisions:0 txqueuelen:1000
         RX bytes:432690788 (412.6 MiB) TX bytes:734404 (717.1 KiB)

Changes

The system board on the node where eth0 disappears was replaced.

Cause

The cause may be that after a system board is replaced the MAC addresses change and need to be reset. You can diagnose if this is the case by checking whether the MAC addresses are the same in the ifcfg-eth0, ifcfg-eth1, ifcfg-eth2, ifcfg-eth3 files in /etc/sysconfig/network-scripts compared with the associated ILOM output in /SYS/MB/NET0, /SYS/MB/NET1, /SYS/MB/NET2, /SYS/MB/NET3 respectively. The MAC addresses should be the same. If they are not they need to be reset.

The steps are:

1. Check the output in /etc/sysconfig/network-scripts for each of:
ifcfg-eth0
ifcfg-eth1
ifcfg-eth2
ifcfg-eth3

For example, output for ifcfg-eth0 looks like below. The field of interest is: HWADDR.

# more /etc/sysconfig/network-scripts/ifcfg-eth0

#### DO NOT REMOVE THESE LINES ####
#### %INITIALIZED FOR BDA% ####
IPADDR=*.*.*.*
NETMASK=255.255.248.0
NETWORK=*.*.*.*.0
BROADCAST=10.128.23.255
HWADDR=00:10:e0:*:*:1c
GATEWAY=*.*.*.*
DEVICE=eth0
BOOTPROTO=static
ONBOOT=yes
HOTPLUG=no
IPV6INIT=no

2. Log into the ILOM for the node and from the CLI check:

# ssh nodex-ilom
-> show /SYS/MB/NET0
-> show /SYS/MB/NET1
-> show /SYS/MB/NET2
-> show /SYS/MB/NET3

For example, output for /SYS/MB/NET0 looks like below. The field of interest is: fru_macaddress.

-> show /SYS/MB/NET0

/SYS/MB/NET0
   Targets:

   Properties:
       type = Network Interface
       ipmi_name = MB/NET0
       fru_description = 10G Ethernet Controller
       fru_manufacturer = INTEL
       fru_part_number = X540
       fru_macaddress = 00:10:e0:*:*:1b
       fault_state = OK
       clear_fault_action = (none)

The MAC addresses in /etc/sysconfig/network-scripts/ifcfg-eth* and /SYS/MB/NET*, where * ranges from 0 to 3 should be the same. In other words the MAC addresses in the HWADDR and fru_macaddress fields should be the same like:

HWADDR=00:10:e0:*:*:1c
fru_macaddress = 00:10:e0:*:*:1c

In the case of the error raised it may be the case that the HWADDR address from /etc/sysconfig/network-scripts/ifcfg-eth0 and fru_macaddress from /SYS/MB/NET0 are not the same. The HWADDR addresses from /etc/sysconfig/network-scripts/ifcfg-eth1,2,3 and fru_macaddress from /SYS/MB/NET1,2,3 should respectively be the same.

Solution

The solution is to do the following as 'root' on the node exhibiting the error:

1. Manually update the files in the /etc/sysconfig/network-scripts to reflect the new MAC addresses as shown from the ILOM.

Note: If you want to back up the /etc/sysconfig/network-scripts/ifcfg-eth0-ifcfg-eth3 files you can make backups in the /etc/sysconfig/network-scripts directory, but you must make sure to prepend the filename. For example, instead of "ifcfg-eth0.bak" the backup file should be "bak.ifcfg-eth0". Otherwise if you want to append to the file name, move the backups to another directory. This is required because all configuration files in the directory /etc/sysconfig/network-scripts/ will be honored by the network service which does not understand which file is the active one and which one is backup. Failure to backup the files as specified here will cause the eth0 interface to fail to be recognized.

2. Reboot the node to bring up the eth0 interface.

If you try to restart the network interface for eth0 with the commands below to turn the interface off and then on,

# ifdown eth0

# ifup eth0

or by running "service network restart", you may find the failures for ifdown eth0 and ifup eth0 respectively:
# ifdown eth0
RTNETLINK answers: No such file or directory
RTNETLINK answers: No such file or directory

# ifup eth0
Device eth0 does not seem to be present, delaying initialization.

This is avoided by rebooting the node to bring up the eth0 interface

Currently there isn't an automated way to do this from the BDA, but in the next release we are adding an API call that should make this easier.

Attachments

This solution has no attachment