I use Fedora 13 (minimal installation), ISC DHCP server 4.1.1-P1. I'm running Fedora 13 too on a separate machine (minimal installation) with ISC DHCPclient 4.1.1-P1. My goal is to do some IPv6 testing. When I use the DHCP client manually (dhclient -6 -timeout 5 -d -v eth0), the client retrieves and installs a dynamically assigned IPv6 address.
Additionally, the client gets 1 DNS resolver address and 1 DNS search list. The resolver address and DNS search list always get written into /etc/resolv.conf. Perfect, exactly what I need. But when I do a "service network restart" or restart my system (which should trigger the same, as far as I understood), the DNS resolver addr and the search list get written into /etc/resolv.conf.save and the /etc/resolv.conf stays as before. How can I change this behavior? I don't need the .save file, I just want to have /etc/resolv.conf replaced by the latest DNS information.
I'm trying to set this open-nic-dns (69.164.208.50) as default (instead of 192.168.2.1 - router-ip). So if I try to edit /etc/resolv.conf to this nameserver - but if i restart it's 192.168.2.1 again.
I just installed F10 from a Live download iso on a Sony VAIO laptop with an Intel PRO/Wireless LAN2100 3B Mini PCI Adapter. The linux driver is ipw2100 and functions without problems.
After the install I accepted the update downloads offer and downloaded somewhere around 170 updates. The kernel version is 2.6.27.19-170.2.35.fc10.i686.
I am using a Linksys WRT55AG as a DHCP server. Using NetworkManger it assigns IP addresses without incident.
My problem is that I can't get name resolution to function without manual intervention.
I can ping other wireless nodes on my network and the response to $ dig -q www.yahoo.com is: ->>HEADER<<- opcode: QUERY, status: REFUSED, id: 30153
The /etc/resolv.conf file contains no entry for 'nameserver=' and if I manually edit /etc/resolv.conf, it is overwritten on reboot or restarting the NetworkManager service. I am aware of /etc/sysconfig/networking/profiles/default/resolv.conf.
Additionally, in the var/log/messages file I can see that NetworkManager makes links between /etc/resolv.conf and the default version.
However, after either rebooting or restarting the NetworkManager service, if I manually enter a DNS address 192.168.1.1 (my wireless router) under the DNS tab of the system-config-network 1.5.95 application. This address does not persist after either rebooting or restarting NetworkManager.
If I disable NetworkManager, I can't get the WRT55AG to assign the wilreless adapter an IP address.
So the bottom line and my question is what do I need to do in order to simply boot and have wireless connectivity?
To what file does system-config-network 1.5.95 write the DNS address entered from it's GUI?
No nameservers found; Try putting DNS servers into your ifcfg files in /etc/sysconfig/network-scripts like so: [...]
After running wvdial. I suppose I have to create a file ifcfg-ppp0, but what does it have to contain in order to automatically grep the DNS servers from wvdial?
In Ubuntu, it works just fine, just that I don't have an Ubuntu installation anymore.
I find it very frustrating that documentation suggests using PEERDNS=no in /etc/sysconfig/network-scripts/ ifcfg-eth0 in order to override resolv.conf information only to find out that both NM and network init scripts ignore DNS1, DNS2 and DOMAIN settings.So my question is how do I correctly configure resolv.conf settings that will persistently override dhcp information on a F12 system that has NM removed and uses the network service scripts?
How can I keep /etc/resolv.conf static ... as in preventing it from being changed by the network manager ... without actually disabling the network manager (so wireless users can get their wireless setup that includes IP address, but leave the resolver configuration as is)?
I need help figuring out why my resolv.conf keeps changing to this, causing me to not be able to access the outside internet and only the local network.I thought I fixed it by getting rid of the loopback interface and adding in the eth0 interface in /etc/network/interfaces URl...And then things would work temporarily, and eventually the nameserver in resolv.conf would get reverted again.
I am using kubuntu 9.10 64-bit on AMD M500 machine i removed networkmanager (because i had frequent disconnect) and installed wicd but the /etc/resolv.conf had a comment on top which says that it is to be configured by networkmanager (still) I put the DNS in there manually and it works What to put in there so that it uses DHCP ? Maybe its set to be not written by anyone other than networkmanager ??
When I boot my computer (Ubuntu 9.10) /etc/resolv.conf is written as:# Generated by Connection Managernameserver 10.33.8.11 10.33.4.136However, if I try to do anything from the command line, I get an error like:host: parse of /etc/resolv.conf failedI then need to edit /etc/resolv.conf to:
# Generated by Connection Manager nameserver 10.33.8.11 nameserver 10.33.4.136
I'm having problems with my resolv.conf file. Every time I reboot my computer it goes blank and I have to re-type it before I can get on the network. How do I prevent it from being overwritten?
I am using KVPNC (recommended by the VPN owner) on Ubuntu 9.10 using my Universities direct Ethernet connection.
I have downloaded all the .ovpn scripts to a directory. I have downloaded KVPNC from the repository (apparently for 9.10 you do not download from the KVPNC website).
I then use sudo kvpnc and select a server (since there are many based on the .ovpn files).I click connect and the script works and tells me I have a connection. See below for an example. All appears to work.
But when I actually load any Internet program e.g. Firefox it cannot find a website.This happens for all servers which have different locations around the world.When I disconnect KVPNC I still have no access. I have to remove the Ethernet cable before it works again.
debug: Connect try requested, profile: Stockholm, type: OpenVPN debug: openvpn: /usr/sbin/openvpn debug: Support for TUN/TAP found (compiled into kernel or kernel module already loaded). debug: Default interface: "eth0". debug: IP address of default interface: "my IP address". code....
I realise there are errors e.g. the /etc/resolv.conf must be a symlink but have tried doing this and still have exactly the same issues.
I'm using 9.10 Ubuntu Karmic 64-bit (but same thing happens on my 32 bit VM on same PC).
My /etc/resolf/conf file is not updated with DNS information on first boot. The DNS information is in the /etc/network/interfaces file.
The trick I have to do every time I reboot the PC is:
Code: sudo ifdown eth0 && sudo ifup eth0
First off, why? Second how can I resolve this issue?
I have three NICs eth0 and 1 are normal LANs connected to the internet, but I choose eth0 as the primary. eth2 is a local private one. the DNS information from eth0 should be in the resolv.conf file but isn't. The file is empty at bootup, except for some comments instructing me not to edit this. After running the above sudo commands the file is properly updated and internet works.
Here's my /etc/network/interfaces file's contents:
Code: auto lo iface lo inet loopback auto eth0 #iface eth0 inet dhcp
My network works before but now it has follow issue, when I ping google.com, I got Quote:
ping: unknown host google.com but my network still works partially because I can still connect to a DNS listed in my /etc/resolv.conf to get host google.com's ip adress by Quote:
host -t a google.com ip.of.one.dns and then I ping the returned IP result 66.249.89.104, the ping works fine. the issue is that I can't connect to Internet, what I can remember is that I ever removed /etc/resolv.conf and this file was created again by network-manager, and I already verified the two dns IPs listed in the resolv.conf are both correct and works fine with command host.
I have an Ubuntu 10.04 server. It gets its IP address by DHCP (the router has a fixed reservation so it always gets the same one). However, it needs its nameserver to be itself, and not the DNS servers given out by the router. How can I make dhcp not overwrite resolve.conf?
(Yes I know I could just set a static IP. But I'm testing stuff out here, and want to know if it can be done without doing that.)
The problem is that the above resolv.conf file doesnt work anymore. The only way that our workstations get name resolution is by commenting out the second and forth lines, which happen to be our secondary DNS. This has got me confused because why would resolv.conf be broken by having a secondary nameserver and search path? Shouldnt it just use the first nameserver unless it fails to resolve? Almost seems like just having the secondary in the file breaks name resolution. I cant ping local hosts or anything on the internet unless I do this:
my rounter causes puppy to write an improper dns entry in resolv.conf. Consequently I have to manually write it, then I have to place a command on it to make sure it doesn't get rewritten on boot, or after a couple of hours use. I had this problem with xubuntu as well.
[URL]
I solved this problem in ubuntu by using the chattr command.
[URL]
Puppy won't let me do this. When I use the chattr +i /etc/resolv.conf, it gives me this error. chattr: Inappropriate ioctl for device while reading flags on /etc/resolv.conf I tried chmod 444 but it had no effect on reboot. resolv.conf was rewritten again. How can I make sure this file does not get rewritten. Why won't chattr work?
Im running several linux boxes, all redat enterprise linux. Versions vary between 4 and 5.
Since we have problems with our DNS servers, Im trying to lower the DNS resolve timeout to the minimum to avoid system hangs etc.
However, it seems that nothing I do have any effect. when Im running time nslookup whatever.com or time host whatever.com while my first DNS server is down, it always takes about 1.09sec as opposed to 0.09sec when the first server in the list is available. I tried playing with options timeout:X attemps:X but it simply doesnt seem to do anything, even if I set it to high values. Some internet posts suggested having timeout:0.5 or less, but that also doesnt seem to do anything.
when I use: options timeout:0.3 attempts:1 - doesn't do anything. options timeout:1 attempts:1 - doesn't do anything. options timeout:10 attempts:3 - doesn't do anything.
It doesn't work when I put it in the first line of the file. It doesn't work when I put it in the last line of the file.
It doesn't work if I put it in "export RES_OPTIONS=options timeout:X attempts:Y" I tried it on 2 different machines.
I'm having an issue with resolv.conf being empty (apart from the resolvconf header) upon first login. I am assigned a DHCP lease fine, it also appears that DNS works briefly as all my NFS file shares seem to be working ok and they have name rather than an IP address in /etc/fstab. I have to manually re establish my lease by using the Network Manager panel each time I first login.
I use wvdial and a 3G modem to connect to the network. On the same host I also run bind9 which is authorative for my local network.The problem is that when wvdial connects, it overwrites /etc/resolv.conf, which means that DNS lookups on that host no longer get made to localhost (as specified by the original resolv.conf file) and end up at the local instance of bind which is authorative for in-house hostnames, but end up with the ISP who doesn't know my in-house host names.The local DNS forwards to the ISP, so the ISP's DNS'es are not required in resolv.conf.How can I prevent wvdial from overwriting /etc/resolv.conf?Removing write privileges from /etc/resolv.conf did not work.
My Internet connection runs slow and jerky, but can be fixed by using the Google public nameservers instead of the default one in this file. Problem is, the changes keep being reverted, even when I chmoded the file to have read-only permissions by everyone.
I use my laptop both at home and at my job. At home it gives me the correct search line.
But when I am at the office I only get 1 search entry while if you boot in windows it gives you 3
(for example):
So this should be in /etc/resolv.conf search domain domain.com my.domain.com
It seems that this is not passed by the (Windows) dhcp server (Active Directory).
Is there any way that linux can detect depending on the network it is in that there are a few search domains added(for example like resolvconf does with ppp tunnels)?
I have resolvconf installed to prevent multiple programs from fighting to update /etc/resolv.conf at the same time an corrupting it.From what I understand resolvconf a package found in the repos is suppose to act as the gatekeeper to /etc/resolv.conf and all programs must go thru it to update resolv.conf. Hence /etc/resolv.conf is written in a orderly manner.
from time to time though something goes awry and resolvconf writes old, wrong stale into /etc/resolv.conf, for a connection that may be down or no longer exists along with the correct info for the current connection.Usually i would fix this by uninstalling and reinstalling the resolvconf package.I was tired of this so i went digging around in
/etc/resolvconf and found the subdirectory /etc/resolvconf/run/interface
This is where it looks like resolvconf caches interface info that goes into
/etc/resolv.conf
I think what happens is resolvconf is suppose to clean this up in an orderly way but sometimes old interface files may be left here.So for instance in mine were the files eth0 and wlan0.with dns info for each stored in the relative file. I no longer have an eth0 only a wlan0 so resolvconf was still pulling info from both
/etc/resolvconf/run/interface/wlan0 and /etc/resolvconf/run/interface/eth0
with the result of an incorrect dns info in /etc/resolv.conf To make a long story short I just deleted /etc/resolvconf/run/interface/eth0 and was good to go. So if you are using resolvconf and run into this problem you may use this work around.
I have a number of CentOS servers with latest 5.xI have 3 nameservers in the resolv.conf files. All 3 nameservers test out fine when checking for domain lookups.I'm having some trouble with ns1 and shut it down.l None of the CentOS machines fail over to the ns2 & ns3 entries. Basically everything stops resolving even though the other 2 nameservers are alive and well.
I had some trouble with the netinst CD and I had to install lenny with the 6DVDs instead. My internet connection only works after I edit the interfaces and resolv.conf files manually. So my question is: can I do that from the netinst CD before it actually needs the connection to install everything? (otherwise I'll just need to use the DVDs.
I have installed Kubuntu 8.04.2 on a USB stick with persistence to keep any changes I make after a reboot and it works fine. I then installed the 185 Nvidia driver to give me higher resolutions and it works fine.
But each time I reboot, my updated xorg.conf is replaced with the default xorg.conf that ships with that version of Kubuntu and a backup is made of my updated xorg.conf (the correct one) which looks like xorg.conf.20100409135913. I have to put the backup xorg.conf back in place to get my Nvidia driver to work with the correct screen resolutions again. Otherwise my screen resolution is too low.
What could be causing this behavor? I'm sure it not the persistence feature of the USB stick failing since a backup is made of my original xorg.conf.
I might add more information. The xorg.conf that gets changed after a reboot says "This file was generated by dexconf, the Debian X Configuration tool, using # values from the debconf database." #
My Debian workstation is on a predominantly Windows network. It is resolving internet addresses (i.e. www.google.com, etc) just fine. However it's not resolving any of the address of machines on the local network and the windows machines do not resolve the linux machines address either. All machines (linux and windows) can access each other just fine via direct IP reference.
My /etc/resolv.conf file has the following IP addresses listed:
127.0.0.1 10.1.1.111 10.1.3.4
10.1.1.111 is a netgear VPN/router that is the primary router for the LAN and maintains a VPN to a remote office 10.1.3.4 is the IP address of the Windows Primary Domain control on which the DNS server is running. That server is at the remote office. The system acts like it is completely ignoring anything I put in the resolv.conf file. From the linux X server I'm using "Network Setting" utility to make changes to the network configuration. This appears to work fine for things like switch between DHCP and static IP.
I'm setting up virtual dev environments and I want the hostname.domain to be something specific, eg- "barney.local". However my /etc/resolv.conf file keeps getting nuked because i'm using DHCP. I placed this into my dhclient.conf:
I tried to setup a dyndns, and ever since I tried to get that setup, I can't access my webmin, or access /etc/hosts/resolv.conf. But I can still access my samba shares I had created before this happened.
