24 comment

  1. Nice topic.

    I faced this problem before, but I solve when I change old mac address with new MAC I get from VMWare inside “70-persistent-net.rules” file.

  2. This can be easily fix with this kind of script :

    # Annoying bug in vmware guest centos6
    # eth0 doesn't exist
    ifconfig eth0 2>/dev/null >/dev/null
    if [ $? -ne 0 ] ; then
            # Rename eth1 with eth0
            echo "UDEV Config..."
            rm /etc/udev/rules.d/70-persistent-net.rules
            # Change ifcfg-eth0 with hostname address (in /etc/hosts)
            echo "Changing eth0 address..."
            mac=`ifconfig eth1 | grep eth1 | awk '{ print $5}'`
            ip=`grep $hostname /etc/hosts|cut -s -f 1`
            rm -f /etc/sysconfig/network-scripts/ifcfg-eth1
            cfg=`grep -v IPADDR /etc/sysconfig/network-scripts/ifcfg-eth0 |grep -v HWADDR`
            echo "$cfg" > /etc/sysconfig/network-scripts/ifcfg-eth0
            echo "HWADDR="$mac"" >> /etc/sysconfig/network-scripts/ifcfg-eth0
            echo "IPADDR="$ip"" >> /etc/sysconfig/network-scripts/ifcfg-eth0
            # Apply changes
            echo "Applying changes..."
            service network stop
            service NetworkManager stop
            ifconfig eth1 down
            udevadm trigger
            udevadm control --reload-rules
            service network start
            service NetworkManager start
    I had it in a .conf file /etc/init/.
    I hope it helps...
    1. It helps for sure…
      Just installed ur script in a template and called by /etc/rc.d/rc.local !
      works perfectly when deploying a new VM. Your script is really helpfull thanks a lot ! great stuff… ๐Ÿ™‚

      BTW Nixcraft when you say “My guess is my network card switched names or renamed eth0 and eth1 as eth1 and eth2.”
      Wrong guess, it’s because of VmWare reassigning new mac address to your clone… then udev create a new rule so eth0 always end as eth1 in your clone.

      Martin Allert go it right about deleting mac address reference in template. It’s the way to workaround and fix this problem.
      I still don’t understand why Redhat still bind mac address to ethernet card … can’t work at all in a virtual environment as hypervisor deciding about hardware stuff!
      And nowadays 70% more Linux servers are virtual… go figure !

    2. Thanks for the help, I tweaked it a bit.

      #! /bin/sh
      # Make sure user running this script is root
      if [ `id -u` != "0" ]
      	echo "You must be root"
      #Checking for rules file
      if [ -e "/etc/udev/rules.d/70-persistent-net.rules" ]
      echo "NetworkManager found"
      echo "NetworkManager not found"
      echo Re-enumerating the network interfaces...
      # Stop networking
      service network stop
      #  ERROR
      # Get the drivers and bounce the NICs
      lspci -nnv | awk '{
      if ($0 ~ /~$/) { getDriver = 0; }
      if (getDriver == 0 && $0 ~ /.*Ethernet controller.*/) { getDriver=1; }
      if (getDriver == 1 && $0 ~ /.*Kernel driver.*/) { printf ("%sn", $5); }
      }' | sort | uniq | while read a; do
      echo Unloading network module $a from kernel space
      rmmod $a
      # Remove the static enumerations
      rm -fv /etc/udev/rules.d/70-persistent-net.rules
      # Remove UUID from ifcfg files
      sed -i /^UUID/d /etc/sysconfig/networking/devices/ifcfg-eth*
      sed -i /^UUID/d /etc/sysconfig/network-scripts/ifcfg-eth*
      # Remove HWADDR from ifcfg files
      sed -i /^HWADDR/d /etc/sysconfig/networking/devices/ifcfg-eth*
      sed -i /^HWADDR/d /etc/sysconfig/network-scripts/ifcfg-eth*
      echo Reloading driver via udev
      # Tickle udev
      /sbin/udevadm trigger
      /sbin/udevadm control --reload-rules
      # Restart networking
      echo "Please wait..."
      service NetworkManager restart
      sleep 10;
      service network restart
      echo "NOTES: 1) If you run system-config-network-tui"
      echo "          Open each NIC and save to automatically add current HWADDR to ifcfg"
      echo "       2) If you run nm-connection-editor"
      echo "          Open each nic and save to automatically add current UUID to ifcfg"
      echo "Network repair is complete..above options are supplemental"
  3. My problem is:
    When I clone a new VM from a template, I get a eth1 despite the fact firstboot is presenting a eth0 for initial configuration.

    The solution is:
    Delete the HWADDR or MAC adresse from /etc/sysconfig/network-scripts/ifcfg-eth0 in the template. After a reboot, firstboot comes up and correctly configures a eth0.

  4. Sweet! Thank you, it actually worked (with the sudo).
    # cd /etc/udev/rules.d
    # cp 70-persistent-net.rules /root/
    # rm 70-persistent-net.rules
    # reboot

  5. I have the same problem, I have done all the steps form the top to reboot,
    but to check when used grep command i could not see other eth ports or MAC address, I think at least I should see 2 mac addresses after reboot. where may i have gone wrong?

  6. Hey,

    thanks it was usefull. But with some bad coding:)


    # cd /etc/udev/rules.d
    # mv 70-persistent-net.rules /root/
    # reboot


    # ifconfig eth1 | awk ‘/HWaddr/ { print $5}’

  7. I perfected the whole process ๐Ÿ˜‰

    # mv /etc/udev/rules.d/70-persistent-net.rules /root/

    Remove MAC-address from the following files
    # nano /etc/sysconfig/network-scripts/ifcfg-eth0
    # nano /etc/sysconfig/network-scripts/ifcfg-eth1

    # reboot

  8. On red hat 5 & 6 it works removing the /etc/udev/rules.d/70-persistent-net.rules files and removing the mac address from the ifcfg-ethX files. This is done on the original machine to prevent the future clone to create a new ifcfg-ethY

    Any Idea on how does it work with suse ?

  9. I was trying to copy a VM that contained a snapshot, evidently there is extra path and file name info in there, which presents a problem because snapshots may be more than 2GB:
    grep -n GameServer1 UDB1-Snapshot1.vmsn
    11495040:nvram = “GameServer1.nvram”
    11495057:displayName = “GameServer1”
    11495058:extendedConfigFile = “GameServer1.vmxf”
    11495074:scsi0:0.fileName = “GameServer1.vmdk”
    11495088:sched.swap.derivedName = “/vmfs/volumes/5321bbd3-ad5bdbf5-9ec4-f01fafd599b9/GameServer1/GameServer1-25db7aa3.vswp”

    I was able to copy this too using the sed utility after inflating the snapshot:
    cat path/GameServer1-Snapshot1.vmsn |sed ‘s/GameServer1/UDB1/g’ > path/UDB1-Snapshot1.vmsn

    This may take a long time depending on the file size. All the other files I had copied normally and edited using vim. Hope this helps.

  10. Excellent post!
    I’m a Linux noob but have to manage and clone some of our vitual Linux servers from time to time – that’s when I faced this issue.
    Played around with system-config-network and added the new ethX adapter (showd up as eth5 in ifconfig) – and that would actually give me network connectivity – but doing it your way also gives the Network Manager full control of the NIC / displays correctly in the GUI.
    Thanks for sharing!


  11. I had the same issue go though different forums but cant fix it..
    I follow a hit and trial ….
    Clone the VM without LAN card .. remove from the edit setting while clonning..
    then after clone i found no lan card in the edit setting and also verify from ifconfig -a
    note the MAC address by adding new nic and change the one at the vi /etc/sysconfig/network-scripts/ifcfg-eth0 by the new MAC address.

    same and assign new ip address from the SETUP and issue resolved…


  12. Delete UUID y MAC from ifcfg-eth0 script

    #rm โ€“f /etc/udev/rules.d/70-persistent-net.rules
    #vim /etc/sysconfig/network-scripts/ifcfg-eth0
    #ip link set eth1 name eth0
    # /etc/init.d/network restart

    Have a question? Post it on our forum!