On 03/04/2016 03:42 PM, Bob Goodwin wrote:
On 03/04/16 18:01, Rick Stevens wrote:
On 03/04/2016 01:38 PM, Bob Goodwin wrote:
I have two Fedora-23 computers running virtual machine manager, this one connects to my ethernet LAN and connects to the internet, the second one, I've just set up connects to the LAN but not the internet, something is wrong with VMM configuration.
Are both of these machines you're talking about acting as hosts for virtual machines?
. Both F23 boxes have vmm running in them. The ifconfig data is from Centos7 in the VM's rather than the host machine.
Oh, then I misread your initial statement. I thought you were talking about the hosts, not the VM guests.
Ok. The first VM (the one that's working) has a reasonable IP address on eth0 (one from the DHCP server part of libvirtd). The second one does not have an IP address on eth0 and that's your problem.
Check the /etc/sysconfig/network-scripts/ifcfg-eth0 file. It should look sorta like this (things like the UUID, HWADDR and DNS stuff will be different):
TYPE="Ethernet" BOOTPROTO="dhcp" DEFROUTE="yes" IPV4_FAILURE_FATAL="no" IPV6INIT="yes" IPV6_AUTOCONF="yes" IPV6_DEFROUTE="yes" IPV6_FAILURE_FATAL="no" NAME="eth0" UUID="e90b53e3-6248-485b-9d52-7388f719c43d" ONBOOT="yes" DNS1=162.249.56.204 DNS2=162.249.56.205 HWADDR=52:54:00:80:03:2D PEERDNS=no PEERROUTES=yes IPV6_PEERDNS=yes IPV6_PEERROUTES=yes
'BOOTPROTO="dhcp"' indicates that it will try to get an IP address from libvirtd's DHCP server. 'ONBOOT="yes"' means that it will try to start the device when booted. If that's set to "no", then just try
ifup eth0
and see if eth0 gets an IP address. If it does, edit the file (as root) and change it to "yes" so it will start on the next reboot.
This one, the working one shows:
[bobg@localhost ~]$ ifconfig eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.122.30 netmask 255.255.255.0 broadcast 192.168.122.255 inet6 fe80::5054:ff:fe65:23a6 prefixlen 64 scopeid 0x20<link> ether 52:54:00:65:23:a6 txqueuelen 1000 (Ethernet) RX packets 162 bytes 11984 (11.7 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 89 bytes 11226 (10.9 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
That is an odd IP address to see on eth0 on a virtual machine _host_. I'd expect a virtual machine _guest_ to have an eth0 address on that 192.168.122 network.
192.168.122 is typically used by the libvirt mechanism to provide IP addresses to virtual machines. libvirt runs a DHCP server offering IPs on the 192.168.122 network to the virtual machines running under it.
Be that as it may, since the IP is on eth0, it's valid and your router is NATting it and allowing you out to the internet.
lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10<host> loop txqueuelen 0 (Local Loopback) RX packets 20 bytes 1716 (1.6 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 20 bytes 1716 (1.6 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
This is what I see in the other one that does not connect:
[bobg@localhost ~]$ ifconfig eth0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 ether 52:54:00:c8:72:84 txqueuelen 1000 (Ethernet) RX packets 39 bytes 2028 (1.9 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
Note that eth0 does NOT have an IP address, so it can't do anything on the LAN (or the Internet).
.
Yes, I understand that, but after following the prompts for installing the centos system in the vm and VMM itself this is what resulted and I don't know what it is supposed to look like. All I could do was compare the working with the non working.
lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10<host> loop txqueuelen 0 (Local Loopback) RX packets 128 bytes 11136 (10.8 KiB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 128 bytes 11136 (10.8 KiB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
virbr0: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 inet 192.168.122.1 netmask 255.255.255.0 broadcast 192.168.122.255 ether 52:54:00:71:23:11 txqueuelen 0 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
virbr0 (virtual bridge zero) is created by libvirt and used to route traffic INSIDE that machine between itself and the virtual machines it's hosting. Note that it has the 192.168.122.1 address (typical of a VM host acting as a router for the virtual machines running under it) and _virtual_ machines running on that host it will have addresses on that 192.168.122 network applied to their eth0 interfaces (assuming they're configured to use DHCP to get an address).
[bobg@localhost ~]$ ping -c 3 8.8.8.8 connect: Network is unreachable
I don't know what I might have done differently on the seond VM install, incidentally both are installed usingthe same centos 7 iso if that is of interest ...
Any suggestion of what to do next will be appreciated,
Give an IP address to eth0 on the machine that's not working and/or leave it using DHCP and ensure you have a DHCP server on the network it's plugged into.
Your network appears a bit weird to me. My LAN uses 192.168.1.0/24 with 192.168.1.1 on the LAN side of my router (the WAN side of the router has a routable public IP). The router also has a DHCP server on it. I prefer to use fixed IPs for my non-mobile devices, so I have the DHCP server restricted to only offer addresses between 192.168.1.192 and 192.168.1.254, letting me use 192.168.1.2 through .191 for fixed IPs.
My first VM host (vmhost-a) has a fixed address of 192.168.1.10 on eth0 (on the LAN). My second VM host (vmhost-b) has a fixed address of 192.168.1.15 on eth0 (also on the LAN). Both of them have default routes that specify 192.168.1.1 (the IP address of my router on the LAN).
When libvirtd starts on both vmhost-a and vmhost-b, it creates a virbr0 interface on each and gives it an address of 192.168.122.1. The thing to note here is that even though both machines have the same address on virbr0, that network is only visible INSIDE the physical machine. This means the 192.168.122 network on vmhost-a does not conflict with the 192.168.122 network on vmhost-b.
If I start a VM on vmhost-b and that VM has its eth0 configured to use DHCP, it will get an address from the DHCP server running as part of libvirtd on vmhost-b. It will get an IP address of 192.168.122.x (we'll use .10 for an example) and a default gateway of 192.168.122.1.
Now, if that VM sends traffic out, it will go through its simulated eth0 (192.168.122.10) and libvirtd will see it on its virbr0 at 192.168.122.1. libvirtd will then route the traffic out its eth0 (192.168.1.15) and the router will pick that up off its 192.168.1.1 interface and route it through to the internet. Replies go back down the same pipeline in reverse. Remember, the router sees the traffic as coming from vmhost-b, so replies go back to it. vmhost-b reroutes the reply traffic through its virbr0 back to the VM.
I hope my example clears it up a bit. There's a lot to absorb there.
Yes, everything helps. My LAN address scheme is about the same as yours. I think I will try assigning a fixed address to the second box, if that works I'll do the same on the one that is working ...
I'll report the result tomorrow'
Thank you,
Bob
- Rick Stevens, Systems Engineer, AllDigital ricks@alldigital.com -
- AIM/Skype: therps2 ICQ: 226437340 Yahoo: origrps2 -
- I will go to my happy place. I WOULD go to my happy place.... -
if I knew where the @$>&$@#* it is! -