|
Alfred,
I would not delete network manager, it would be better if you stop it "service NetworkManager stop" and disable from booting "chkconfig NetworkManager off"....it can turn a useful tool for troubleshooting in the future. In a second note about the multi-NIC, i would focus in the actual card that have the connections....here at work we have system with 8 NIC for redundant links, but my main focus is in the actual ports that have connectivity. Take a look at /etc/sysconfig/network-scripts.....focus in the ifcfg-ethx that is actually working in your box. Don't forget that you can always use system-config-network-tui as an alternative to using the GUI, as long as you have it install "yum install system-config-network-tui -y".....I think you are in the right path, let us know how it turn out. In reference to the host name, take a look at the file /etc/sysconfig/network and /etc/sysconfig/network-scripts/ifcfg_ethx to assign host name information....I hope I didn't confuse you more....the Redhat documentation is rocksolid, take a glance at it. Best of luck to you.
>>> Alfred von Campe <alfred@xxxxxxxxxxxxx> 8/17/2011 9:50 AM >>> On Aug 16, 2011, at 18:37, Tom H wrote: > If you mean after the install, what's the output of "chkconfig --list > NetworkManager", "chkconfig --list network", and your NIC's ifcfg-X? I ended up re-installing the system from DVD this morning (don't have my kickstart server set up yet), and this time I noticed a "Configure Network" button in one of the screens. It was there that I enabled a setting to start the network on boot, and now it's working "out of the box". I can't stop to wonder why upstream decided to make this optional. Who doesn't want their network up, especially because it's not obvious how to start it once you are logged in? But the story doesn't stop there. When I enabled the network during the installation, the connection name was "System eth0". When I logged in (as root, because this is just a test system and I hadn't configured local or network accounts yet), the network was running but using a connection named "Wired connection 1" and I had two ifcfg-X files: # cat network-scripts/ifcfg-eth0 DEVICE=eth0 HWADDR=bc:30:5b:e7:a7:1c NM_CONTROLLED=yes _ONBOOT_=no BOOTPROTO=dhcp TYPE=Ethernet USERCTL=no PEERDNS=yes IPV6INIT=no #cat network-scripts/ifcfg-Wired_connection_1 HWADDR=BC:30:5B:E7:A7:1C TYPE=Ethernet BOOTPROTO=dhcp DEFROUTE=yes PEERDNS=yes PEERROUTES=yes IPV4_FAILURE_FATAL=yes IPV6INIT=no NAME="Wired connection 1" UUID=44c68048-52ba-4e5e-807f-adf8a881552f _ONBOOT_=yes LAST_CONNECT=1313586374 Why on earth would NM create another config file for the same interface? Also, how do I tell NM to include the DHCP_HOSTNAME variable in the config file so that my DHCP server creates a DNS entry? I think I don't like the NM "overhead" in CentOS 6. Is it just a matter of doing a "chkconfig --del NetworkManager" to get rid off it? I'll be working on getting the CentOS 6 installation automated via kickstart later this week and hope to resolve all these issues before then. Thanks for all the help/pointers, Alfred _______________________________________________ CentOS mailing list CentOS@xxxxxxxxxx http://lists.centos.org/mailman/listinfo/centos |
_______________________________________________ CentOS mailing list CentOS@xxxxxxxxxx http://lists.centos.org/mailman/listinfo/centos
