Controllerless Networks

I configured a IAP-Mesh-Bridge with two IAP315 to bridge some networks from one site to another.

The bridge itselt is up and runnning. However, when I try to ping the switch on the Mesh-Point-Side I face a packetloss around 45%. The notebooks on the Mesh-Point-Side are running fine. No issues with them.

My assumption is that they are communication all the time. The switch is not communicating all the time and I can force to get the switch back to 100% response when I flash ARP tables. Otherwise I have to wait some time until the switch is doing any communication and the ping is back with full response. After some time it drops again.

Main issue is if we connect phone on the Point-Side, I will not be able to call them all the time.

Any suggestions why this is happening?

Is this indoor our outdoor?

What is your channel Width for that mesh (20mhz, 40 mhz, 80 mhz)?

If this is outdoors, you would have better luck with a 20mhz width channel.

You should type "show ap mesh link" on both sides to see the RSSI (rate) and tx/rx rates.

 http://www.arubanetworks.com/techdocs/Instant_423_WebHelp/InstantWebHelp.htm#CLI_commands/show_ap_mesh_link.htm

Hi,

we have an RSSI from 37 on the local and 39 on the remote side.

It is using 80 Mhz channel. Do you think it is more stable to use 20 Mhz?

Yes, it is used for outdoor meshing.

20 mhz is more stable, yes.   If you can uncheck "wide" and "80 mhz" in the arm profile, during a maintenance window, you should.  

We configured 40 Mhz for the moment (I will suggest to use 20Mhz) but the issue is still the same.

I would assume that it is not the wireless connection but the port configuration or some other special configurations that the clients that are not sending traffic all the time to time out. Are there any options like that?

What is the distance between mesh APS?

The distance is about 10-15m.

The RSSI in 40Mhz chanel width is 41 (from my point of view a very good value).

I fear we need to have a look at port configurations and access roles.

mneesen,

Did you ever find the solution to the problem you described? I have a similar situation with an Aruba IAP wireless bridge setup. I have always on devices at the site that work perfectly fine but my switch management IP and a UPS with a management IP drop alot of packets. I have to setup a ip sla continuous ping on my switch to keep the traffic from dropping to it. Would be wonderful to resolve so I can monitor my UPS for power outages but had to disable since I kept getting up/down alerts due to dropped packets.

Jeremy

hi Jeremy

there is a known issue with IAP mesh, where a topology like this:

sw1 --- mesh portal --//    //-- mesh point --- sw2

will have packet loss when trying to ping sw2 from the sw1 network side. This is apparently fixed in 8.4.x software due to some rearchitecture of the mesh code. I have asked R&D for some more info about it, has something to do with idle time which I think matches your observvaton of needing a continuous ping.

What version are you on and could you consider an upgrade to 8.4.x at some point soon ?

-jeff

Hello Jeff,

We are running 8.3.0.5 on our IAP bridge using 374 AP's. I see an 8.4.0.0 version available to download under standard release. Will this version include the new mesh setup? I'm willing to try it since my setup has minimal usage at this time. Thanks.

Jeremy

Hello Jeff,

I went ahead and updated my bridge setup to 8.4.0.0. I removed my ip sla continuous ping on my switch with no dropping of packets so far. Also have added my UPS back to monitoring. No drops on polling snmp variables to it. Seems to have resolved the issue. Thanks for you help.

Jeremy

Thank you for testing.

It would be good to have a faster solution for the next case...