Weird networking issue (ping and other service related)

I asked this on the IRC channel a couple of days ago and didn’t get the
issue resolved I am hoping that I would get a better answer here. I have
two machines, my main box which is running openSUSE 11.1 and my home server
which is running openSUSE 11.2 milestone 1 (was testing a problem with my
network card with 11.1 and wanted to see if it was in 11.2, but got this
solved with help from this forum).

As for the problem both machines are connected to a Linksys router and both
machines can get out to the internet, tested by pinging Google and having
zypper check for updates. It gets weird when I try and access the machines
through the LAN on the 11.2 box it can talk to my router and my 11.1 box
just fine, but when I try and talk to the 11.2 box (it hosts my NFS, and I
have apache on it while I learn PHP and to test my CGI scripts) from my 11.1
box, or from the ping utility in my Linksys router, I get errors about how
the destination is unreachable.

Anyone have any idea of what is going on, or did I stumble on some kind of
time paradox between the two SUSE versions?

Do you have firewall enabled on these machines?
If so, disable them and see OR open the required ports in the firewall.

syampillai wrote:

>
> Do you have firewall enabled on these machines?
> If so, disable them and see OR open the required ports in the firewall.
>
>

I should have mentioned that I already tried that and even with the
firewalls off the router and 11.1 box can not ping the 11.2 box

Can you post the output of the following commands from both the machines?

/sbin/ifconfig
/sbin/route

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Agreed, verify the subnet mask and default gateway just in case they are
hosed. Preferably:

ip addr
ip route

Good luck.

syampillai wrote:
> Can you post the output of the following commands from both the
> machines?
>
> Code:
> --------------------
> /sbin/ifconfig
> /sbin/route
> --------------------
>
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
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=y8m7
-----END PGP SIGNATURE-----

ab@novell.com wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Agreed, verify the subnet mask and default gateway just in case they are
> hosed. Preferably:
>
> ip addr
> ip route
>
> Good luck.
>

Here is from my 11.1 box:

Big-O:/home/vendion # ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 brd 127.255.255.255 scope host lo
inet 127.0.0.2/8 brd 127.255.255.255 scope host secondary lo
2: ra0: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast state DOWN qlen 1000
link/ether 00:22:5f:1f:67:5c brd ff:ff:ff:ff:ff:ff
3: eth0: <BROADCAST,MULTICAST,PROMISC,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast
state UP qlen 1000
link/ether 00:1d:72:a6:a5:fe brd ff:ff:ff:ff:ff:ff
5: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state
UNKNOWN
link/ether 00:1d:72:a6:a5:fe brd ff:ff:ff:ff:ff:ff
inet 192.168.1.101/24 brd 192.168.1.255 scope global br0

Big-O:/home/vendion # ip route
192.168.1.0/24 dev br0 proto kernel scope link src 192.168.1.101
127.0.0.0/8 dev lo scope link
default via 192.168.1.1 dev br0

Here is from my 11.2 box:

SEServer:/home/vendion # ip addr
1: lo: <LOOPBACK, UP, LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 brd 127.255.255.255 scope host lo
inet 127.0.0.2/8 brd 127.255.255.255 scope host secondary lo
2: eth0: <BROADCAST, MULTICAST, UP, LOWER_UP> mtu 1500 qdisk pfifo_fast
state UP qlen 1000
link/ether 00:08:a1:03:74:35 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.100/24 brd 192.168.1.255 scope global eth0

SEServer:/home/vendion # ip route
192.168.1.0/24 dev eth0 proto kernel scope link src 192.168.1.100
169.254.0.0/16 dev eth0 scope link
127.0.0.0/8 dev lo scope link
default via 192.168.1.1 dev eth0

Hmm, why have you got a bridge interface on your 11.1 box? Have you been playing around with VMs?

Assumming you don’t need the bridge (?)
Try the following in a terminal (as root/sudo) on your 11.1 box:

# /sbin/brctl delif br0 eth0
# /sbin/ifconfig br0 down
# /sbin/brctl delbr br0 
# /sbin/ifconfig eth0 inet 192.168.1.101 netmask 255.255.255.0
# /sbin/route add default gw 192.168.1.1

Then try to ping in both ways.

‘brctl’ is in the bridge-utils package. So install it if it is not there.

If you’re using VirtualBox, get the latest (non OSE) version. It will take care of the bridge. But you might have a good reason to need a bridge that I don’t know.

please try again wrote:

>
> Assumming you don’t need the bridge (?)
> Try the following in a terminal (as root/sudo) on your 11.1 box:
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> # /sbin/brctl delif br0 eth0
> # /sbin/ifconfig br0 down
> # /sbin/brctl delbr br0
> # /sbin/ifconfig eth0 inet 192.168.1.101 netmask 255.255.255.0
> # /sbin/route add default gw 192.168.1.1
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> Then try to ping in both ways.
>
> ‘brctl’ is in the bridge-utils package. So install it if it is not
> there.
>
> If you’re using VirtualBox, get the latest (non OSE) version. It will
> take care of the bridge. But you might have a good reason to need a
> bridge that I don’t know.
>
>

I have the bridge because I use KVM and without it I wasn’t able to get the
guest OS to have Internet access. Reading the KVM wiki the bridge for the
network card seemed like the best solution and that is why I choose to use
it.

Adam Jimerson wrote:

> please try again wrote:
>
>>
>> Assumming you don’t need the bridge (?)
>> Try the following in a terminal (as root/sudo) on your 11.1 box:
>>
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>> # /sbin/brctl delif br0 eth0
>> # /sbin/ifconfig br0 down
>> # /sbin/brctl delbr br0
>> # /sbin/ifconfig eth0 inet 192.168.1.101 netmask 255.255.255.0
>> # /sbin/route add default gw 192.168.1.1
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>> Then try to ping in both ways.
>>
>> ‘brctl’ is in the bridge-utils package. So install it if it is not
>> there.
>>
>> If you’re using VirtualBox, get the latest (non OSE) version. It will
>> take care of the bridge. But you might have a good reason to need a
>> bridge that I don’t know.
>>
>>
>
> I have the bridge because I use KVM and without it I wasn’t able to get
> the
> guest OS to have Internet access. Reading the KVM wiki the bridge for the
> network card seemed like the best solution and that is why I choose to use
> it.

I should also add that if I take down my bridge connection and bring my
wireless online, ra0, the problem still exists so I’m sure this is not
related to the bridge.

A bridge is the best choice if you want host and guest to be both accessible from each other. However nat would be safer if the guest only need Internet access and never shares its filesystem to the host. But it is pure theory since I don’t know KVM. Can you not boot without KVM for the surpose of testing your “real” lan first? Then you shouldn’t have a bridge interface.

According to the output you posted, you’re bridging from eth0, the bridge is alive and has the IP that eth0 should have in order to talk to your lan. Also I don’t understand why the bridge interface has the same ethernet address than your NIC … I would call that “a half bridge”. But again, I don’t know how KVM organizes things.

Anyway, if you reboot without KVM and assign an IP in your local network range to eth0 (any IP starting with “192.168.1” other than 192.168.1.1, your router and 192.168.1.100, which appears to be your other box) you will be able to talk to your lan through that device.

hmmm … does your router run DHCP ? Maybe that’s the problem (just guessing). Your router assigns an IP to the bridge and not to your wired interface (?). Please reboot with KVM disabled if you want to investigate that problem.

Then, to display the current tcp/ip settings of your eth0, type:

ifconfig eth0

You should see something like that (except for the Hwaddr) :

eth0 Link encap:Ethernet HWaddr 00:18:F8:0F:7B:EF
inet addr:192.168.1.101 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: fe80::218:f8ff:fe0f:7bef/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:342923 errors:0 dropped:116 overruns:0 frame:0
TX packets:242311 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:451152533 (430.2 Mb) TX bytes:22012169 (20.9 Mb)
Interrupt:16 Base address:0xc000

If the inet addr line is missing or shows an ip not in your local network range, you won’t be able do ping any machine in your lan. That’s why I told you how to disable the bridge and assign a static IP to that device in my first post.

Because KVM is just a kernel module I’m not sure how to reboot with out it short of uninstalling KVM and related packages. What doesn’t make sense to me is when I kill my wired connection and just use my wireless, which is not bridged, can not ping the 11.2 machine but I can take down the bridge on eth0 and try without that just to be sure. The thing that has got me is when the 11.2 box was running 11.0 I did not have this problem, I was able to talk to it even with the bridge.

Just bringing up a device, whether cabled or wireless, is not enough to get connected to anything. That device should receive an IP in the range of the network it belongs to.
According to your output, none of your network devices has got an appropriate IP. It is not suprising that they are not able to ping others in your lan.
I’m sure somebody here can tell you how to configure your network devices on openSUSE with “user friendly” tools (like network manager or similar stuff, that I never use)

As for the kvm module, I don’t think it is essential, and renaming the module before the next reboot should do the trick. You can rename it back later. I don’t know how it is called, but something like kvm.ko somewhere in /lib/modules/… sounds logical.

What do you mean they don’t have an apporiate ip? My Linksys router is running DHCP and is giving out IP of 192.168.1.101-192.168.1.150 (192.168.1.100 is a static IP because it is my home server) both machines are in the same network 192.168.1.xxx and have the same subnet. None of these machines are running Network Manager, the 11.2 box has a static IP and the 11.1 machine is using DHCP from the router, what is wrong with this confutation? Also it is only the 11.2 machine that can’t be pinged by other machines, the 11.2 machine can ping my 11.1 box and my router and both machines are able to get out on the Internet. This was not an issue when the 11.2 machine was running 11.0, my network configuration was the same as it is now, the only difference is I installed 11.2 on the server instead of 11.0 which has only a year or so of support left.

Now I ask why would disabling my virtual machine system make more of a difference than the bridged connection that I have set up?

Try to manually assign an IP over 150 to eth0 and see if you can ping then.

ifconfig eth0 inet 192.168.1.151 netmask 255.255.255.0

And could you please post the ouput of ifconfig -a ?

please try again wrote:

>
> vendion;1987445 Wrote:
>> What do you mean they don’t have an apporiate ip? My Linksys router is
>> running DHCP and is giving out IP of 192.168.1.101-192.168.1.150
>> (192.168.1.100 is a static IP because it is my home server) both
>> machines are in the same network 192.168.1.xxx and have the same subnet.
>>
>
> Try to manually assign an IP over 150 to eth0 and see if you can ping
> then.
>
> # ifconfig eth0 inet 192.168.1.151 netmask 255.255.255.0
>
> And could you please post the ouput of ifconfig -a ?
>
>

On which machine the 11.1 or the 11.2 box?

The 11.1. The other one looks fine. You don’t have problem with the 11.2 box, don’t you ?

You won’t break anything by manually assigning an IP to a device. The worst that could happen is that you lose your connection to the lan, but since you don’t have a connection at all … Such a setting won’t survive a reboot anyway. It is juste for the purpose of testing to see if you’re able to ping your lan under normal circumstances ( = without bridge ). Whether or nor the bridge causes the problem, it certainly doesn’t help to get closer to the solution.

I cannot help with the wireless device. I’ve been doing that only once under Unix and I don’t remember the details. It’s also a little bit different: Linux has a separate command called “iwconfig” to handle wireless devices. “man iwconfig” might give you some info. Basically, you need to know or scan the channel and give the right (Wep or WPa2) key to the access point. Be aware if your router acts as a wireless access point running DHCP and offering 150 IPs, that 149 of your neighbors can use your connection for whatever purpose they need. Not to mention that it has possibly already happened and somebody managed to reconfigure your router while you’re still looking for misconfigurations … >:) Anyway, it it advisable not to run a wireless DHCP without a good reason anf even so to assign static IPs whenever possible.

A very simple test you could do, which doesn’t require any knowledge or effort, would be to boot the box which is apparently not connected to the lan from a Linux live CD. I don’t know if openSUSE has one. Ubuntu, Mandriva and Knoppix do (allthough avoid Mandriva. It detects certain netcards wrong) It will bring up Gnome, probably detect your NIC and load the appropriate module, then send a DHCP request. If you cannot connect at this point, you have a problem with your router or the cable. If you can connect, that’s a problem with your openSUSE. Also I’m not sure if it makes sense to have two network interfaces, both sending DHCP requests (?).

I really would :

  1. disable the bridge
  2. forget about the wireless interface (for now)
  3. see if I can connect to the lan with eth0 configured as DHCP client

If such a simple configuration doesn’t work, a more sophisticated one, with bridging and more interfaces - won’t either.

Sorry guys, yes of course they do! I’m downloading Milestone1 Gnome liveCD and give it a try.