Heeeeeelp! Wireless issues for months now!

[CyBeRkId2002]

Hi all,

 

I am hoping someone can help as I have absolutely exhausted all of my knowledge regarding wireless networking and am pulling my hair out!

 

Since we came back from the Summer Holidays we have had intermittent issues with our wireless network grinding to a halt and stopping computers from functioning for minutes at a time. It takes new students logging in 10-20minutes to log in. It seems to make those who are logged in grind to a halt and the whole computer run slowly, even thought the task manager looks perfectly normal.

 

Changes over summer

 

New bank of laptops

New edge switches for a new block of the school (HP and our core network is Cisco)

 

What have I tried to diagnose the issue

 

Checked event viewer of affected computers

Disabled various AP's in the worst hit areas - nothing to report

Replaced AP's in affected areas with new Ubiquiti boxes

Created a new VLAN and put select laptops on it

Disabled auto-updates to affected computers

Temporarily power-down edge switches to see if that has any affect (tricky as this now powers a whole new wing to the school)

 

I am absolutely running out of ideas! I cannot reproduce the problem so don't even know where to begin to find a solution... any help would be greatly appreciated.

 

Michael

[dry]

Have you had a look at your DC? Do you have any wired clients you could jump on to when the issue starts, to see if it is actually your wireless network or a problem with your domain controller?

 

What is your wireless system? You mention Ubiquiti but does that mean you've got Unifi all over your site?

[mikeprice]

You have probably already eliminated this but anyway

 

 

we had a similar issue a while ago where an access point decided that it would reset its config at random. It inevitably set itself to a channel very close to another access point - hence messing everything up.

A similar problem also exists here at the far end of the building. We have no computers or network points done there but if someone takes a laptop the chances of the wireless working are 50:50 as it is so close to the houses at the back (only about 10 yards!)

The access points in the houses frequently interfere with the school wireless

[Blue_Cookeh]

Does this only happen when you have the new batch of laptops online? How many clients are sitting on each AP?

 

You may need to turn power down on the APs and put more in to load balance a bit.

 

Where are you routing your VLANs?

[CyBeRkId2002]

Blue_Cookeh: Nope... seems to now happen regardless of which batch of laptops are switched on. Quite possible on the AP's, hadn't really considered interference before but now that is seeming quite sensible. VLAN's are routed at either of our two core switches. Have looked and looked at this but as far as I can tell all seems fine.

 

dry: DC seems fine. Desktops are completely unaffected by this problem and makes me think it could be VLAN routing. Our main wireless network is Cisco. We have had some new ubiquiti but these problems occured before we even plugged in a single Ubiquiti AP. Tried switching an area to Ubiquiti to resolve the issue but doesn't seem to have worked.

 

mikeprice: Possibly channel related and something I am now going to look into.

[Blue_Cookeh]

How many APs are within close proximity of where the issue is happening? It sounds awfully like you've got too many devices on an AP. How many are using a single AP at a time?

[Wave9_Andy]

A few things to consider after asking my technical colleagues....

 

In no particular order

 

1. has it ever been surveyed to see what’s happening with AP placement, channel allocation and surrounding Wi-Fi interference?

2. check with tools such as WiSpy to see what is happening on the RF spectrum – non Wi-Fi interference

3. what models are the cisco / unfi AP’s and what standards are they running -802.11 a/b/g/n/ac?

4. how many clients are associated to a particular AP at any one time?

5. are clients rate limited?

6. what applications are being used is there any commonality when issues are experienced?

7. check the controller configuration parameters and logs – are AP’s disconnecting?

8. understand and check network infrastructure – what is doing the routing and is there latency and loss to it. i.e. use a tool like ping plotter to ping the AP from the normal vlan. Does any packet loss coincide with issues seen on the WLAN?

9. Are there any points of contention in the network – i.e. internet connection?

 

Andy

[psydii]

Further to comments around chanenels here is a great blog about it.

 

Channels 1, 6 and 11 | MetaGeek Blog

 

Don't forget that you *will* see multiple APs on the same channel because you have a large deployment. What you need to see is that there is at least 20dB between the loudest and next loudest AP at any given location.

 

 

 

Also use wireshark to check what a given client is seeing at layer 2/3. Perhaps there is something flooding the air. If you have a macbook pro you can see all wifi packets, not just those for your computer.

[CyBeRkId2002]

Ok networking experts... I have setup a laptop on Linux and managed to grab some preliminary network scans from when the network grinds to a halt (which is every lesson at the minute). The only thing that seems strange is that whenever we have a network slowdown we have a really high number of packets going to all-hsrp-routers (and usually from one or two laptops, but these vary and it isn't always the same machines). Could be perfectly normal, just something I noticed that was different when problems occur.

 

Are there any network specialists out therte that I could send our grabs to who could have a nosey at what could possibly be going on. Have had a look over our HSRP config and from my admiteddly wooly knowledge on the subject everything seems absolutely fine.

[psydii]

I would suggest that you should not see any HSRP traffic on the air, and since that is the first piece of quantitive hard evidence of a potential source of the problem.. let's run with that for a while:

 

If your routers are failing over, or at least doing some sort of re-calc to determine which one should be active, depending on your network configuration you might have found the 'smoking gun'.

 

I have no experience of HSRP, but I would be tempted to block it from going airside. I would also expect that there to be a configuration/design flaw in the network somewhere. Perhaps some sort of loop caused by the wireless (perhaps two routers can see each other twice via their wired interfaces and through a wireless SSID?), however...

 

Given the changes over the summer I would suspect that the new HP kit is seeing the Cisco Propriatry protocols on the wire and just forwarding them on, and it may be the place to look would be at the edge of the cisco kit where it connects to the HP network.

 

Once the rogue packets get into the wireless network, depending on your precise configuration, you may even be seeing these packets from APs connected to previously well behaving Cisco kit.