5

I have two tomato routers. One is bridged wirelessly with the other.

I have a new server on the network. It's running Ubuntu Server 11.04.

They are all connected like this:

A - Linux PC
B - New Server
C - Mac Mini
D - Macbook
T1 - Tomato 1
T2 - Tomato 2

They are connected like so:

A -----+-T1  ==== wireless bridge ==== T2----- ADSL modem
       |                               | C & D Connected wirelessly to T2
B -----+

A, C & D do not experience any issues.

I have an active SSH session to B from A and it's not experiencing any loss.

B, the new server occasionally cannot ping T2 and therefore cannot connect to the internet. However, A can always contact B and B can ping A and B When the network is lost, B can still ping T1, but not T2 yet at the same as B has lost connection to T2, A can still ping T2.

Any ideas on what this could be? there is nothing that gives any clues in any of the logs on either router or the linux server.

One thing that is interesting is that I set up a ping running between B and T2. T2 has the IP address 192.68.1.1

Here is what I am seeing:

From 192.168.1.1 icmp_seq=26 Destination Host Unreachable
From 192.168.1.1 icmp_seq=27 Destination Host Unreachable
From 192.168.1.1 icmp_seq=28 Destination Host Unreachable
From 192.168.1.1 icmp_seq=29 Destination Host Unreachable
From 192.168.1.1 icmp_seq=30 Destination Host Unreachable
From 192.168.1.1 icmp_seq=31 Destination Host Unreachable
From 192.168.1.1 icmp_seq=33 Destination Host Unreachable
From 192.168.1.1 icmp_seq=34 Destination Host Unreachable
From 192.168.1.1 icmp_seq=35 Destination Host Unreachable
64 bytes from 192.168.1.1: icmp_req=36 ttl=63 time=3.40 ms
64 bytes from 192.168.1.1: icmp_req=37 ttl=63 time=5.70 ms
64 bytes from 192.168.1.1: icmp_req=38 ttl=63 time=2.25 ms
64 bytes from 192.168.1.1: icmp_req=39 ttl=63 time=2.18 ms
64 bytes from 192.168.1.1: icmp_req=40 ttl=63 time=3.12 ms
64 bytes from 192.168.1.1: icmp_req=41 ttl=63 time=2.15 ms
64 bytes from 192.168.1.1: icmp_req=42 ttl=63 time=1.97 ms
64 bytes from 192.168.1.1: icmp_req=43 ttl=63 time=

And it cycles to being reachable and not.

So I guess you could say the question is, why is the router responding that it cannot be reached?

hookenz
  • 4,043
  • 3
  • 32
  • 46
  • I've just disabled STP (Spanning Tree Protocol) on T2. So far so good. Any thoughts on that? could STP be the cause? – hookenz Aug 05 '11 at 02:51
  • What does your IP scheme look like for all the devices involved? Destination Host Unreachable is a response a hop will return when the next hop is not found - it is because you have reached the destination network, but the route to the host specified does not exist. This could be a routing problem or it could be a connection problem. – MaQleod Aug 05 '11 at 05:10
  • @MaQleod - local network is 192.168.1.x. T1 & T2 are all on the same network address range It's very weird. It's only this machine. Even reinstalling it has the same issue. – hookenz Aug 22 '11 at 22:27
  • 4
    These are the classic symptoms of trying to bridge behind a client WiFi connection. T2 only has one client because there is only one radio talking to it. Thus it cannot see both A and B as its client. An access point is prohibited by the WiFi specs from putting traffic on the air if it isn't to any of its clients (unless you're using WDS). If you use WDS, everything will probably work perfectly. – David Schwartz Mar 01 '13 at 14:41
  • @DavidSchwartz do you have a source for that? I am interested in learning more about it. – prateek61 Apr 02 '13 at 12:04
  • 1
    @prateek61 Start with this page on [WDS](http://en.wikipedia.org/wiki/Wireless_distribution_system). – David Schwartz Apr 02 '13 at 18:24
  • check the routing table of the two routers and show us what you found... an incorrect static route or an incorrect network specified in RIP could cause this problem – Mortie May 21 '12 at 15:31
  • David, I really feel like this may be the correct answer because he is using only a single subnet. – George Spiceland Feb 23 '14 at 05:20
  • Interesting. I've not got this setup anymore so it's no longer a problem. And I've since changed my wifi routers. Although since then I've had other weird issues occasionally with wifi where devices couldn't see each other when they were meant to be able to. Sometimes they would and sometimes wouldn't. The solution was to turn on proxy arp and all the issues went away. – hookenz Feb 23 '14 at 22:29

2 Answers2

2

“Destination Host Unreachable” on the same network segment indicates that resolving the IP address to a MAC address using ARP has failed.

Because wireless links assume that there is only one device at the client end, bridges (Repeaters too, btw) cannot work over WiFi unless using special means like relayd or WDS. Usually, you can’t even bridge wireless client interfaces to Ethernet interfaces. There is only one exception to this: A special Broadcom wireless driver that performs “MAC Masquerading” of sorts.

WDS, on the other hand, has to be supported on both access point and client: In this configuration, a wireless packet carries 4 MAC addresses instead of just two. In addition to the immediate sender and recipient addresses, there’s also the source and destination address. This way, true Ethernet bridging is possible.

WDS does have one problem though: Usually, it limits encryption to WPA, which is insecure. Atheros devised a proprietary extension to WPS that allows WPA2. As such, there is but one solution for stable Ethernet bridging over WiFi: Using Atheros devices on both ends.

STP doesn’t have anything to do with this.

The OpenWrt wiki also has a very good article on client mode wireless.

Daniel B
  • 60,360
  • 9
  • 122
  • 163
0

Since it's a wireless bridge, I would suspect a flaky connection. STP would not be the real cause of that. It's purpose is to prevent bridging loops so that your network doesn't become saturated with looping packets and become unusable. Turning it off is usually a bad idea as it now won't protect you from that situation.

if the port is "bouncing" STP will block the interface while it re-elects a root bridge. It will also block the interface if it detects a bridge loop. I suspect there is more to your topology than you are showing in your diagram. Does the new server have another interface?

Keith
  • 8,013
  • 1
  • 31
  • 34
  • Nope. There is not more to the topology than shown. That's everything. Since turning off STP it's been stable. – hookenz Aug 05 '11 at 03:28
  • The connection between T1 & T2 is very stable, not flaky. I've had no problems with host A communicating over the bridge, just B. – hookenz Aug 05 '11 at 03:29
  • What are their MAC addresses? – Keith Aug 05 '11 at 04:46
  • Ok, the STP didn't fix it. Note: B is the only maching having issues. how does knowing the mac address help? what are you thinking? – hookenz Aug 05 '11 at 04:50
  • Ok, nm. I was wondering if there was a duplicate MAC that might confuse STP. But that seems to not be the problem now. Is an interesting problem. Can you configure the router to send syslog messages to a syslog host (say set up on A)? – Keith Aug 05 '11 at 08:14