CentOS 5 Networking :: Dosen't Enslave Bonded Interfaces On Boot?
Feb 9, 2011
I have a rack of IBM x335s. They are all running 802.1Q over 802.3ad, that is VLAN trunking over two physical Ethernet bonded with LACP. Its been working fine on all of the servers for years.One server recently had its hard drives upgraded and CentOS reinstalled. It is a minimal install. The server has been reconfigured exactly as it was and has been verified and compared against the other working servers.It will not enslave the interfaces on boot. You can manually enslave and then bring up the network, so that much of the configuration is verified. It simply refuses to enslave... which fails everything that comes after thatHere is a snippet of the message log on the machine that does not enslave on boot
Jan 31 22:28:57 vm3srvp01 kernel: NET: Registered protocol family 10
Jan 31 22:28:57 vm3srvp01 kernel: lo: Disabled Privacy Extensions
Jan 31 22:28:57 vm3srvp01 kernel: IPv6 over IPv4 tunneling driver
I'm having trouble getting Fedora 15 to work with my em/bond/br configuration as it did in Fedora 14. I've got a bonding interface over em1 and em2, and then individual bond vlans (bond0.2, bond0.10, etc...), I then have a similarly named bridge interface (br0.2, br0.10, etc...) for kvm. Regardless of if I assign bond0.2 or br0.2 the IP address, I am unable to ping the gateway, but the native untagged VLAN (bond0, br0) has no issue at all.
To try to isolate this issue down to the switch or Fedora 15, I went with a plain em setup and configured em1, and em1.2 which works as expected. This issue has been plaguing me since Fedora 15 came out and I'm about to trash it and go back to Fedora 14 (Which didn't have this issue), but some of the new kvm features have kept me trying to get over this incomprehensible hump.
I have a series of four config scripts which all run in sequence of each other.first occurs during the whole kcikstart thing.Kickstart then tells rc.local to run step2, which then tells rc.local to run step 3. No steps are repeated, after they are run they delete themselves from rc.local.After each step, it reboots.Step 3 is responsible for creating the /etc/sysconfig/network-scripts/ifcfg-eth1 file, and populating it. It greps ifconfig -a for the mac address of USB0.
#! /bin/bash
# Check to see if the usb nic takes eth2, if it does, move it to eth1[code]............
I have two eth interfaces bridged in CentOS, one of these interface is connected to a Windows Server, I can do a ifconfig down on both these interfaces and there is no change of status at the windows side, I don't get the message that 'network cable has been unplugged'
Is this behavior normal ? I think when you shut an interface than you get the interface down message at other end ? And what becomes the status of the bridge if I shutdown one of its interface ?
I have 4 NICs and am trying to create 2 bonds. The first bond is mode 1, the second is mode 4. What happens though is makes both bonds mode 1. Here's my modeprobe.conf:
alias bond0 bonding options bond0 miimon=80 mode=1 alias bond1 bonding options bond1 miimon=100 mode=4 lacp_rate=1
I have a weird issue that I have not seen on any forum. My jaunty on DELL studio laptop seems connected to net, but I can not access any network service (ssh, firefox etc.). But when I connect a cable the cable lights blink as it should be and in wireless connection my wifi light blinks.
It was working 2 days ago without problem, and I have not done big changes recently.I removed and reinstalled network-manager and network-manager-gnome. Nothing changed. I see a message in each restart as follows (when Openafs is starting). I can reproduce it with "/etc/init.d/openafs-client restart"
Code:
ADVISEADDR:error in specifying interfaces: no existing ip interfaces found
So I keep reading that bonding ethernet devices is supposed to be easy. I have followed several tutorials and tried a few things on my own, and I can't seem to get it right. I currently have a setup that allows me to ping internally on my network, but when I try to ping externally I get the error: connect: Network is unreachable
Before I bonded the interface my connection worked fine. Here are my config files. I am running centos 5.3 on a virtual machine. /etc/sysconfig/network-scripts/ifcfg-bond0
which file can setup up my network interfaces? i was able to setup the nameserver using /etc/resolv.conf but I cannot find anywhere to configure the ip and and gateway. I could't find /etc/network/interfaces as you do on Debian
Now that IPv6 is becoming more and more common, I found the need to disable IPv6 on some interfaces but have it enabled on other. I found that /proc/sys/net/ipv6/conf/*/disable_ipv6 does exactly that. I am now wondering if anybody knows, why are networking scripts so counterintuitive. /etc/sysconfig/network has an option:
NETWORKING_IPV6=yes
All this option does is disable some ipv6 services (dhcpv6...), it does not disable IPv6 in whole (as one would assume...that's why you had to disable it with module parameters). Searching for more IP6 related config option, one can find that /etc/sysconfig/network-scripts/ifcfg-* scripts can contain:
IPV6INIT=yes
Again, this option does not disable IPv6 protocol on the interface, it just skips running ifup-ipv6/ifdown-ipv6 scripts. I added a /sbin/ifup-pre-local. Now, this script runs before ethX entries are created (other scripts run when it's already too late) in /proc directory, so it modifies default values which are then used after those entries are created:
I have 3 Interfaces for a different LAN's and when I start one interface the another interfaces goes down.How can it's possible?I configure my ethernets as:
If I try to add a new interface (eth1) to /etc/network/interfaces, I get
Code: * Reconfiguring network interfaces... SIOCSIFADDR: No such device eth1: ERROR while getting interface flags: No such device SIOCSIFNETMASK: No such device
[Code]...
How do I add 2 interfaces and get anyone of them to work, as available ?
Just something that struck me while working on our virtual servers today.
I have bonded 3 NICs at the host in Ubuntu Server 8.04 LTS. They are using mode 0 for Round-robin. Point is to increase the speed/performance of all the servers, but mainly the fileserver. The fileserver is a virtual server running Ubuntu Server 8.04 LTS on VMware Server 2.0.
1) I noticed the NIC in the slave OS reported link speed as 1000 and Im unable to change it as the NIC (virtual one) doesnt support it. Does this not really matter, as the NIC doesnt exist, and it will run at higher speeds anyway? Or do I have to remove the bond on the host, bridge all 3 interfaces from the host to the slave OS, and then make a bond in the slave OS?
2) While at it, does mode 0 only increase performance on data being sent from the host or does it also increase the available incoming bandwidth?
I'm curious; for a current project a server has 4 network interfaces (eth0 and eth2 are in a bonded pair using the bonding module, and then eth1 is being used for management, while eth3 is currently unused but may be utilized for iscsi in the future).
We need to configure two different hostnames, one for the bonded pair (bond0) and one for eth1. Now I have been googling around to try and figure out how the heck this is done but I can't find much of anything. In some methods it mentions using the hostname command, however this isn't persistant after reboot.
Another mentions just editing the /etc/hosts file but I thought that was only used locally? Both bond0 and eth1 will be on separate VLANs with different DNS servers, so it needs to end up with two different FQDN's
we use for the installation of our machines bladelogic. We have different servers. Some servers have only one network interface, but it can be 2, 4 or may be more. There is always one network device for using PXE, but it is not always eth0.Is there any way to run kickstart without the entering of the PXE-Device so, that kickstart checks all the network devices in the system?
after using Gnome on my 11.3 system I wanted to go back to KDE4. So I installed the patterns and now I can choose it from gdm and it works fine. But if I also try to use kdm, nothing works. I see the kdm login screen without any users. Autologin isn't set in system settings. Any idea how to reset all the kdm settings to default? KDM Screenshot (oxygen theme):
For some time now it's not loading properly, just get a big blue screen with a title bar. This is using 10.04 64bit with Firefox 3.6.15.I think it's since I upgraded to 10.04.
I found that the ATI driver doesn't work on Opensuse 11.3, which is quite annoying!but what they discussed doesn't work for my graphic card, which is ATI FirePro MV 2260. The successful test on Opensuse 11.3 is Radeon cards. I always got 'kernel not match' error.Is there anyone have the same problem like me?
My new Squeeze machine detects all of my hardware including the LAN and WLAN cards, but only brings up the LAN card. There isn't even an entry for it in /etc/network/interfaces for me to modify. I tried adding the info manually but then the entire network wouldn't start. I had it configured as a static IP and I know the commands were correct due to them working on the LAN card. I believe that the name "wlan0" was causing the problem, but how else can I bring it up at boot?
Recently (I can't put an exact date on it but its been within the last few weeks) on boot I've had to wait two or three minutes for my system to bring up my network interface, in a way its never done before...The output of /var/log/boot.log is as follows:
Code: Welcome to Fedora Press 'I' to enter interactive startup.
I'm on an embedded system that doesn't have Gnome, and I'm trying to startup networking automatically using /etc/network/interfaces. Here's what I have.
[Code]....
eth0 comes up just fine. wlan0 comes up, but it's unable to acquire a DHCP address. I added the following lines to /etc/rc.local, and wlan0 comes up all the way, but I'm not too crazy about this hack.
I noticed the debian was not giving me sound playback to some things, such as some games. but other games worked.
So being the ignorant idiot i am i blindly ran alsaconf and did something and now my sound dose not work at all, but if i go to the top panel, click on system- go to preffrences then sound, i can select my input out put device and I happen to use a Usb audio device with a mic.
If i play the test sound i actually get the continuous beep that plays, i can hear it.but now, the Mic dose not work for things like Skype or anything else. also, the sound dose not work when i load ANY games. I come from a windows background if that helps.
when i type the code:
I get a message that says: Rockwell International Device 4310
If i'm not a complete fool, then i believe that is my sound card.
I think that after running "alsaconf" a second time i was able to mount the correct sound device. but it gives me the error message:Terminating processes: 3721.
For adjustment of volumes, use your favorite mixer.
So, is there a way to reinstall my alsa drivers or something. sound dose not work after this, by the way. And will updateing debian lenny solve this? i am not sure which version i have, but it is for sure the 32-bit version. im at my wits end. Reinstalling the whole os can be an option, but i really dont want to do that. first i would have to reinstall Windows xp and install debian useing a windows installer. booting from an ios dose not seem to work for my system. I have an old computer, it was built for windows 98 computer.
I want to dual boot Windows Server 2008 r2 and Ubuntu 10.10 . First, I installed Server 2008 r2 then Ubuntu. After the installation, Grub only found Ubuntu, and "Windows Recovery Enviroment (loader)" on dev/sda1. The Windows OS is installed on dev/sda2. When i load "Windows Recovery Enviroment (loader)" it does a chdsk, then reboot the PC.
How can i make grub find the Windows OS?
Edit: The partition on dev/sda1 is called System Reserved
I upgraded Ubuntu version 8 to version 9.10 and VMware dosen't work no more. After I launch WVware Workstation a message appears: "Before you can run VMware Workstation, several modules must be compiled and loaded into running kernel." I select install (yes), and after that it's giving me error: "Unable to build kernel module. See log file /tmp/vmware-root/setup-30979.log for details."
I'm testing some network setups, and bought two dirt-cheap USB network interfaces. My plan was to attach these two to an embedded device (sheevaplug) and mess with firewall setups. However, the two interfaces are absolutely identical; they have the same MAC, and lsusb -v is the same for both, line by line (except for device number, which tells the order the devices are plugged in, I believe?). Also, this is a single usb bus (single port, with a hub).
Now, obviously I can't make much of a firewall setup if I can't tell the interfaces apart. I realize I probably hit a brick wall, but does anyone have thoughts on this?
I have two interfaces, gprs0 and gprs1, both connected to the Internet from the same HW device, but through two different access points I just read another blog entry with a guy who had two network cards, using one as a back up - I guess this is a little similar, but should be much easier I am guessing - I am just lost Using ifconfig (IP address given by the network) and route to setup the connection,
ifconfig gprs0 10.20.30.40 up route add default dev gprs0 ping 66.102.7.99 works fine afterwards
[code]....
There is data being transmitted and received to device again, I guess the kernel just doesn't know where the incoming data should go without the correct routing tables? Or what exactly is going on?The question I have is how do I configure the device to do the correct routing, using only the ifconfig and route commands? Maybe too trivial for you guys, but I am getting tired of goofing around not knowing what is up and down anymore
I am trying to build a linux distro based on xubuntu that I will install in my car. I need a window manager thats fast and dosen't have to have the minimize, maximize, and close buttons at the top. I also have somewhat of an idea on how to change from xfwm4 but not 100%
I installed the latest release of Kubuntu on to my dell laptop about 3 days ago, I fell asleep last night and woke up to my laptop not being connected to my wireless(it was when I fell asleep, it didnt reboot or anything overnight an dno one touched it), so I plugged in an ethernet cable and it still did not detect an internet connection, then finally it connected through my phones mobile internet via Usb.
However none of the built in networking devices are working, I have them all enabled.My WLAN Interface says "Unmanaged" and Networking interface complains the cable is unplugged even when it is not.I know you guys will need some additional info from me to help me.I just used my mobile broadband to install Wicd network manager but thats just a temporary fix, as it doesnt fix my ethernet issue and I want to be able to use the network manager itself to manage my networks.
