Cloud Admin Club

When you install CentOS, RHEL, or OEL with graphical Interface, it comes with NetworkManager. The NetworkManager daemon maks networking configuration and operation easy and automatic as possible by managing network interfaces, either Ethernet, WiFi, or Mobile Broadband devices. Besides, NetworkManager provides a rich API to control network settings and operations. However, when configuring servers It’s preferable to have their networking settings to be consistent. therefore, This article explains how to disable NetworkManager, hence, configure the network and set IP ethernet manually.

The following steps are suitable for versions 6 and 7 of RHEL, OEL, and CentOS.

Firstly, Identify which card you want to configure by executing the ifconfig command with parameter -s to shortlist the output.

[root@s59 ~]# ifconfig -s
Iface    MTU    RX-OK RX-ERR RX-DRP RX-OVR    TX-OK TX-ERR TX-DRP TX-OVR Flg 
eno1     1500  4596894    0   5617 0        41057    0      0      0 BMRU 
lo       65536     440    0      0 0        440      0      0      0 LRU 
virbr0   1500        0    0      0 0          0      0      0      0 BMU 
[root@s59 ~]#

In view of the output of our example, you can see that the server has three network interfaces, as follow:

eno1	LAN Network interface
lo	Loopback interface
virbr0	Wireless Network interface

In our case, LAN Interface eno1 is our target.

Disable NetworkManager

Firstly, Stop NetworkManager by stopping its service.

systemctl stop NetworkManager.service

Secondly, disable Network Manager using the following command.

systemctl disable NetworkManager.service

manually configure network setting.

Initially, navigate into the network configuration directory. then, edit the configuration file for the ethernet. For example, if the card name eno1 then the filename is ifcfg-eno1.

vi / etc/sysconfig/network-scripts/ifcfg-eno1

After that, add the following parameters in the file if they are not exist

ONBOOT=yes 
NM_CONTROLLED=no 
BOOTPROTO=none 
IPADDR=XXX.XXX.XXX.XXX
NETMASK=255.255.255.0
# you can use PREFIX=24 instead of NETMASK=255.255.255.0
GATEWAY=<GATE WAY IP XXX.XXX.XXX.XXX>

manually configure network in OEL 7, CentOS 7 - ifcfg-eno1

Optional you can add the following lines to force DNS on specific Card

 DNS1=<DNS IP XXX.XXX.XXX.XXX>                
 DNS2=<DNS IP XXX.XXX.XXX.XXX>

Manually configure DNS resolve file.

make sure it has a reliable DNS IPs, for example, google public DNS 8.8.8.8

 vi /etc/resolv.conf  
 nameserver 8.8.8.8

manually configure network in OEL 7, CentOS 7 - resolv

Finally, type :wq to save & exit the editor.

Restart network services

For version 6, use the service command.

# Service restart network

On RHEL 7 use systemD command, though, executing the above command will redirect to systemD command

# systemctl restart network.service

NFS allows you to share local disk over your local network. NFS originally supported by Linux, however, in recent releases of Windows supports it. Sometimes, you might face the error message mount.nfs: mount system call failed or check lease failed on NFSv4 server when you mount shared directory from NFS server. hence, failing to mount your directory.

Investigating the issue

During investigating any issue you cloud obtain knowledge about many things more than getting a direct solution from someone. hence, I suggest continuing reading.

Check connectivity to the server.

The first thing is to confirm that you don’t have any connectivity issues to the server. because sometimes network guys do unintentionally.

you can check nfs status by executing rpcinfo command.

[root@MFVMDML-UNOCONV01 var]# rpcinfo -p 172.16.10.26
program vers proto port service
100000 2 tcp 111 portmapper
100000 2 udp 111 portmapper
100024 1 udp 58355 status
100024 1 tcp 34069 status
100003 2 tcp 2049 nfs 
100227 2 tcp 2049 nfs_acl  
100021 1 tcp 55381 nlockmgr
100005 1 udp 56628 mountd
100005 1 tcp 58276 mountd

The above results reported by the server saying I’m OK, however, it’s a misleading report. therefore, further investigation is required.

Check client OS logs.

Further scrutiny is required, hence, we will dive into OS logs and check kernel ring buffer messages.

Our objective to search for our famous error message “mount.nfs: mount system call failed” or error related to NFS. So, execute dmesg command as shown below.

[root@MFVMDML-UNOCONV01 ~]#dmesg
[1981456.923442] nfs: server 172.16.10.26 not responding, timed out
[1981456.923533] nfs: server 172.16.10.26 not responding, still trying
[1981637.403831] nfs: server 172.16.10.26 not responding, timed out
[2118031.199289] NFS: state manager: check lease failed on NFSv4 server 172.16.10.26 with error 13
[2118036.225762] NFS: state manager: check lease failed on NFSv4 server 172.16.10.26 with error 13

Similar results might show that the NFS server is not responding. So that leads to focus our efforts on the server

dmesg is used to examine or control the kernel ring buffer.

Solving the problem

All NFS servers maintain a table of local directories that are accessible to NFS clients. Yet, restarting NFS services solve nothing. therefore, We need to force NFS to re-check its shares table and update it with the current configurations. luckily NFS tools come with one called exportfs.

This tool maintains table of exported NFS file systems as described in man pages.

login into NFS server with superuser privileges. then, execute exportfs command with parameter -a to rebuild all exports tables.

root@mfvmmdls01:/var/www# exportfs -a

Network

manually configure network in OEL 7, CentOS 7