2 weeks ago, our core layer 3 switch (3Com 5500G EI) and DHCP server (Windows Server 2008 R2) were accidentally powered off. I turned them back on, checked the configurations, and did not see anything indicating any configuration was lost.
A few days later, our users started reporting problems receiving IP addresses on their wireless devices. (Looking at the vWLAN logs, it looks like the problem started after the switch and DHCP server were turned off, but the problem was not widespread yet.) I looked at our vWLAN controller and all devices connecting to our guest SSIDs/VLANs were showing a 0.0.0.0 IP address. Devices on the main SSISs/VLANs (Native AP VLAN) were still working.
I rebooted the DHCP server that afternoon, but that did not help. That night, I rebooted the 3Com switch and that seemed to get the guests SSIDs/VLANs functioning normally again.
Last week, both wired and wireless clients (main and guest VLANs) were experiencing trouble receiving an IP addresses from our DHCP server. This led me to believe something happened to our 3Com switch when it was powered off, so I swapped it with another 3Com 5500G EI. I exported the configuration from the original switch to the replacement switch and so far, our wired clients seem to be back to normal.
On the wireless side, though, some devices connecting to the guest SSIDs/VLANs are still showing a 0.0.0.0 IP address in the client tab. It is primarily devices connecting to SSIDs that use web-based LDAP authentication, although an occasional MAC address authenticated device will show 0.0.0.0 as well.
All our locations are active in vWLAN and I can ping the default gateway for all the locations from vWLAN. I am running the most recent version of vWLAN and firmware for the 1800 and 1930 WAPs.
Can anyone shed light on why some clients are still receiving a 0.0.0.0 IP address?
see if you can get a packet capture with wireshark on one of the clients that is unable to get an ip address, also does the clients mac address show up in dhcp address leases or even the windows event log?
Have you checked the access points to verify that the correct locations are available to the ap's. When the access point boots it tries to find a dhcp server in each location (vLAN) and if it cannot then you will see the issues you are describing. The clients are able to authenticate but do not have access to that vLAN so they cannot get an IP.
Is it possible that when the core switch was powered off that the ap's rebooted because of the control channel timeout, if they did then the access points may have booted but were unable to detect a dhcp server in the appropriate vLAN's. If this is the case then a reboot of the ap's would re-establish a connection to the correct vLAN's so long as the core switch was functioning.
No, I do not think any information from the clients ever reaches the Windows Server. I do not see any information show up in the DHCP console, Event Viewer or the DHCP log files (C:\Windows\System32\dhcp) for any of the MAC addresses that show 0.0.0.0 as their IP address.
In vWLAN click on the Status tab, then click on Access Points, now expand the Locations for each access point to verify the correct Locations (vLAN's) exist.
The WAPs are set to DHCP and their controller access mode is set to discover (I use DHCP Option 43). I know the WAPs are talking to the DHCP server as I can reboot them and they come back every time. When we were running on the old 3Com switch, they would not come back online.
Every WAP is seeing the locations it needs to see. Just to test, I used putty to get into a few WAPs and I was able to ping the DHCP server and the default gateway for each subnet they need to see.
If your APs show the locations under status like Justin indicated then check the role the clients are in to make sure DHCP is allowed or all: For a client to get an IP address - DHCP (or all traffic) must be allowed outgoing, and DHCP server (or all traffic) must be allowed incoming.
Im not familiar with the 3Com switches but do you have a dhcp helper entry in your config to help clients find the dhcp server, if it was needed?
Also a test would be to take a pc and put it in the same location or vlan as your wireless clients and see if an ip addresses is issued.
Yes, I have a DHCP helper set on all the VLANs. I can tag the Ethernet NIC on a computer with either the main or guest VLAN and receive an IP address on either subnet via DHCP.
Just to give everyone some background, I have used the Bluesocket vWLAN and WAPs for 4 years and have had the same configuration in place during that time (switches, VLANs, DHCP server, etc.). Up until the power problem 2 weeks ago, everyone connecting to the guest network would always receive a 10.252.x.x IP address. I have not changed any configuration on our switches or the vWLAN software.
I noticed this morning a client connected to the 1930 WAP in our building initially showed a 0.0.0.0 IP address when it connected, but after a refresh of the clients tab, it showed a 10.252.19.5 IP address, like it should.
I also noticed a co-worker’s Android phone that did receive an IP address (I saw the lease in the DHCP console), but was still showing 0.0.0.0 in the vWLAN clients tab.