Desperate for help Ruckus controller restarted after changing management VLAN

[Dantech]

Hi

 

I spent the last two days logging the MAC addresses, IP addresses and Port Numbers that all our our 36 active Ruckus access points are joined to in order to complete the work started last summer. I had everything working but on the default VLAN, VLAN1. The AP's were on VLAN 1 and so was the Ruckus 3000 Controller. I followed the help instructions and configured the access points to VLAN 8 which is the IP address range I want my clients to be issued an address in. I entered 8 in the Management VLAN in "Access Point Policies" and I entered 8 in configure system/device IP settings for the controller as well. After restarting the controller I cannot ping it at all.

 

I have every port that the AP's are plugged into untagged on the default VLAN (1) and tagged on VLAN 8 (WIFI). The Zone Director was untagged in VLAN1 and Tagged in VLAN 8.

 

I cannot ping on the fixed address assigned to the controller. I even tried changing the port to untagged on the default VLAN and restarting the box.

 

I took my fluke tester and connected it directly to the Zone Controller whilst disconnected from the network and I got no IP address.

 

Foolishly it dod not even occur to me to back up the config because I did not expect there to be a problem.

 

Can anyone help - I'm holiday tomorrow and have to have it for lessons on Monday!

 

Mark

[robk]

Is anything giving out dhcp on the new vlan?

 

You may be able to gt to a command line terminal via the console port on the controller, this would at least let you see the IP address the controller is using to be able to get back in and save, change your config.

[nicholab]

Does it have the old ip address?

 

Have you used wireshark to sniff just connected to the zone director?

 

Also I would start with untagged on vlan1 as the Zone Director only needs to be on vlan 8 if you are tunnelling connections.

[Dantech]

Is anything giving out dhcp on the new vlan?

 

You may be able to gt to a command line terminal via the console port on the controller, this would at least let you see the IP address the controller is using to be able to get back in and save, change your config.

 

Yes I have a dhcp helper address from my Layer 3 core switch. Fortunately I have an in-line fluke tester and was able to see it had retained the original address which was an address in VLAN1 (default) but was Native to VLAN8. I got help from the good guys at c-p-ltd who we bought the kit from and have a contract with, but in the end I decided it was quicker to set the controller back to the default since all my ap's would retain their settings, but just get a new IP address. Because the old address was in the wrong VLAN the only way would be via the console cable (and I have never done this and can't find it!)

[Dantech]

Does it have the old ip address?

 

Have you used wireshark to sniff just connected to the zone director?

 

Also I would start with untagged on vlan1 as the Zone Director only needs to be on vlan 8 if you are tunnelling connections.

 

Hi nicholab, thanks for that. I was advised to use wireshark, but am a bit new to it. I had forgotten I could just plug a laptop straight into the ZD and then I saw how useful wire shark was. It showed me the address of the ZD and that it was trying o reach my DNS server to resolve an address. You were dead right about leaving it on vlan1 - I don't understand tunnelling yet and messing with setups when most of the lessons rely on wireless is not something I do lightly! I had to put a change request in just to make this change (has to be formally in writing with a plan saying what I would do if it goes wrong).

 

Once I reprogrammed it again and left the ZD in vLan1, all the AP's started to come back, but this time because I had tagged all their ports to VLAN 8 (WIFI) and untagged them on the default they all came up with a WIFI VLAN address as they should. Interestingly because some of my ap's have power injectors, some have netgear unmanaged 10/100 switches and some have decent HP PoE switches, many of them needed the power restarting to rejoin the network. I finally left at 7:10 with only 5 I could not get restarted because the buildings were locked and inaccessible. I found one little trick though - on the HP PoE switches I just disabled the PoE and a few seconds later turned it back on. It didn't work everywhere and those ones will need a hard reset.

 

Hope my experience is a lesson to others to TAKE A BACKUP! If I had backed up the config I could have had it back in ten minutes. I suppose because there was no mention of this in the help file and because I did not think it would happen it did not occur to me. Still, I now have all my AP's in the right VLAN and therefore my clients will get the WiFi addresses and therefore the traffic is properly segregated which is the whole point.