Cisco 1200 AP reassociation issue.

[mikespenard]

Network layout is as follows

router 10.0.0.1

AP1 10.0.0.2 (cisco 1300 in root ap mode)

BR1 10.0.0.3 (cisco 1300 in workgrp bridge mode)

AP2 10.0.0.4 (cisco 1200 in root ap mode)

[router]--ethernet1--[AP1]--wireless1--[BR1]--ethernet2--[AP2]--wireless2--[wifi client]

ISSUE:

When the wifi client first associates to AP2

everything works fine. I can ping all devices from

the client, and visa versa.

If the client reassociates to AP2 it can ping

AP2 and BR1, but not past BR1, it can no longer

reach 10.0.0.2 and .1

Power cycling BR1 fixes the problem.

Also, any other node parallel to BR1,(on the workgroup lan,ethernet2) has no issue, it can always ping any node.

Attached is sh ver and sh run for each device.

[dixho]

This is a well known issue on the BR1310. Please upgrade the BR1310 to 12.3(2)JA software.

[mikespenard]

My 2 1300's are 12.3(2)JA

The 1200 however is not. Does the known

issue effect this model as well?

My .txt files have #sh ver info.

[dixho]

Sorry. I should have looked at the text files before respoding to the post.

The AP1200 is not affected by the bug. We need to work on this a little bit. When you experience the problem, please do a show bridge on all the devices. Find out if there is anything wrong with the bridge table (i.e. learning a MAC address in the wrong interface)

[mikespenard]

Note1: The 1200 has been upgraded to 12.3

Still have the same issue.

Note2: I tried replacing the 1200 with a linksys wap11.

Still have the same issue.

Note3: doing "#clear bridge" on BR1 gets the node working again(until it re reassociates, then its broken again)

Note4: If you look at #sh bridge on AP1 when it is working there is a entry for the Wireless Client .66a1

0002.2dbf.66a1 forward Vi0 P 472 219

As soon as Wireless Client reassociates, this entry

is gone. If i do #clear bridge on BR1 it comes back

Note5: When the wireless client reassociates and

loses connectivity to AP1&Router, and i try to ping Router(10.0.0.1) I can see APR requests on BR1, but

no replies. (see file)

------------------

MAC INFO

0012.4335.68c0 0012.016c.0ae2 AP1

000f.f858.1a06 0012.4329.6bd0 AP2

0012.016c.0afa 0012.016c.0afa BR1

00-10-60-5C-41-60 Wired client

00-02-2D-BF-66-A1 Wireless client (10.0.0.21)

0040.63d3.dfe2 router (10.0.0.1)

---------------------

Output of "show bridge" on all units, in good and

bad states and other info is attached

[dixho]

The problem description is very close to an existing bug fixes in 12.3(2)JA. As you already install 12.3(2)JA on BR1 already, it looks like a new bug.

I am busy at the moment and do not have time to recreate the problem. If you can wait, I will do it, but I cannot promise a time frame.

What do you mean by re-associate?

If you do not want a workaround quick, configure AP1 as root-bridge and br1 as non-root bridge.

[mikespenard]

I just noticed AP1 is set to be a infrastucture client.

Would that cause this issue?

[dixho]

You do not need it, but I am not sure if this causes problem. AP1 is an AP, and it is not a client of any kind. On top of it, I do not see how it can mess up BR1's bridge table.

[mikespenard]

Ya that didnt fix it.

[mikespenard]

>What do you mean by re-associate?

-Either change to another SSID/network and come back

on.

-Go out of range and come back.

-Pull PCMCIA card out and put back in.

so, associate, de-associate, associate to the same ssid.

>If you do not want a workaround quick, configure AP1 as root-bridge and br1 as non-root bridge.

Are you saying that IS a quick workaround?

Just to make sure what would the 2 station-role settings be.

[mikespenard]

Setting AP1's station-role to bridge root

and BR1's station role to bridge non-root

FIXED the issue.

So this is a valid work around.

If i could be kept in the loop with Cisco if this

bug is worked on and/or fixed it would be appreciated.

Could you clairify to me the differances between

all of the different station types. I cant seem

to find any docs that go into much detail.

My understating is that a wifi client can only associate to an access point and not to a bridge,

but this does not seem to be the case with the 1300's

as it is with other bridges such as a linksys wet11.

I am also confused as to the differances between

root-bridge, nonroot-bridge and workgroup-bridge.

A root bridge is connected to the primary network

segment, and a nonroot and workgroup are not.

But what is happening at the datalink layer?

I take it there must be some difference in the

way bridging is implemented at that layer.

Does a workgroup bridge only support bridging of

802.3 to 802.11, and 1300's in rootbridge nonroot-bridge 802.11 to 802.3 or 802.11 ?

[dixho]

Root bridge is a superset of AP. On top of fucntion provided by AP, root bridge can associate to non-root bridge and run spanning tree as well.

The major difference between a non-root bridge and workgroup bridge is distance. IEEE put a restriction of 1 mile in the distance between any two wireless devices. Cisco implements propietary code to loose the restriction.

There is no mechanism for us to keep you in the bug fix. I wonder if you can open a TAC service request.