Has anyone ran into the problem of creating a network loop when a laptop is connected to both the wireless network and wired?
If so any suggestions on how to fix such an issue? I'm hoping disabling portfast on the port the AP is connected to on the switch will help, but I don't have much faith in it.
CWAP's try your best! =)
Yeah that didn't help at all, still losing a lot of packets...
Could you please give more information on your problem ? Do you connect to the same subnet or different subnets/networks via both wired and wireless ? Are you losing packets on wired or wirelss network ?
My notebook connects to a (next-door) customer network indirectly via wired network ( the notebook connects to my company router which then connects to the customer 's LAN ) and directly via wireless network ( the AP is inside their LAN ). I do not experience any problem in routing as well as lost packets.
The short answer is no, I have not.
However, if you configure your laptop to bridge between two network interfaces (why would you go out of your way to do that?), and the wired and wireless LANs were bridged to each other, you could create a bridge loop not blocked by Spanning Tree Protocol (STP). But the result would be not simply lost packets but complete network failure as multicast packets are forwarded incessantly. The solution would be to turn off bridging in the laptop.
More likely your laptop is multi-homed on either one or two data-links and has an inadequate IP default gateway for your purposes.
Question time: What are you trying to do? What IP address does each interface end up with. Are they provided by DHCP or entered by hand? What is the default gateway? What does each of the two LANs lead to?
I hope this helps. Could you add your location to your forum profile? Thanks. /criss
You got it. The concern is users that bridge their connections. It crushes the network. The real problem lies in how to stop it if they choose to do so? Not so much that it just 'happens' by accident. But let's say for example you have 1200 AP's and thousands of potential users...and one does decide to bridge their connections, what then?
If the physical loop includes a Spanning Tree Protocol (STP) root bridge, and all the other bridging devices in the loop support STP -- such as laptop, access point, Ethernet concentrator(s) -- then STP will block appropriate port(s) and save the day.
In this context "support STP" means the device either actively sources and sinks Bridge Protocol Data Units (BPDUs) and participates in the election of a root bridge, or the device passively forwards BPDUs and does not participate in the election of a root bridge. Either way STP in the STP active bridges will have enough information to block appropriate ports.
If the device filters BPDUs then STP could be prevented from saving the day. It is conceivable that some access point vendor would do such a thing.
It would make a good experiment to set up the scenario and use Ethernet and WLAN protocol analyzers to examine BPDUs on each bridged segment.
I hope this helps. Thanks. /criss
What platform are the laptops running? I looked around a little and for Windows I only find references to bridging functionality in XP.
Microsoft says that XP implements Spanning Tree, there is no configuration for Spanning Tree, and it is enabled by default.
Thanks for the replies/suggestions...to answer some questions...
Yes the clients are Win XP machines.
STP is enabled on all switches/routers(all Cisco, AP's are Aruba). By your post I'd assume laptops have STP enabled as well.
If the XP laptop bridges their wired & wireless connecion the entire network, wired & wireless gets thumped. So according to Criss's post, the AP's would then need STP enabled as well, correct?
Guess I'll have to get my hands on a analyzer. Any suggestions?
Are you sure it is a STP issue??? Run a local copy of ethereal off each interface on your laptop and see if you are seeing any BPDUs sent out the ethernet multicast address.
Some onther things. If both adapters have IP addresses then chances are the L3 demarcation of each segement will prevent any STP type issue.
Another thing is if you are bridging interfaces on XP, IF XP participates in STP one of the other bridges will get the BPDU and elect a root and designated bridge so the loop should be closed at one point if there is one. One interface will be forwarding and maybe the other blocking. So if the AP does not have STP on its ethernet port enabled the upstream switch connected to it should by default.
Also, if you did have a bridge loop you would not drop packets but see duplicates of broadcasts and encounter a performance problem from the traffic levels being high from repeated broadcasts circulating the loop.
Just some things to consider. Never heard of this on a bridge machine before but then again never really use or see anyone use the bridge interface feature.
Let us know how you make out inquiring minds want to know:)