This topic has been marked solved and closed to new posts due to inactivity. We hope you'll join the conversation by posting to an open topic or starting a new one.
WC7600v2 only connects to one switch / strange behaviour
Hi Netgear Community,
We have a strange situation here at my site. We currently use a WC7600v2 with the latest firmware (6.5.1.11) and a wireless network of 32 AP's of various makes and models of the Netgear family.
Yesterday we decided to change over our core infrastructure that our WC7600v2 connected to. The core comprised of the following,
1 x GSM7224P ProSafe 24-port Gigabit L2+ Managed Switch with PoE+,
This switch provided Ethernet connectivity to all the devices (the servers / WC7600v2 connects to this) in the core rack and had a single Ethernet connection the M4100-12GF ProSafe 12-port.
1 x M4100-12GF ProSafe 12-port Gigabit Fiber L2+ Managed Switch with PoE+
The M4100-12G serves as the switch that aggregates all the backbone runs throughout the site at the centre of the star / network and is connected to the GSM7224P via single Ethernet link.
We decided to start the phased replacement of these devices to a fully stacked L3 H3C / HP 10GB stacked switches and decided start with replacing the GSM7224P with a 24 port H3C/HP 5120EI (using the latest 5120.EI-4210G-4510G_5.20.R2222P01 firmware). We took great care in ensuring all devices where using the latest firmware available and began the process. We replicated the settings from each port of the GSM7224P across to the 5120.
Everything went as expected, all servers and routers where connected to the 5120 with no issues BUT when we connected the WC7600v2 the controller was unable to contact any of the devices on the management vlan (vlan 400). We decided to set an ACCESS PORT on the 5120EI to VLAN 400 to see if we could contact the controller from a workstation that appeared directly on the same VLAN 400 on the same switch (the 5120EI).
We knew something was wrong when we were unable to get a DHCP address from the DHCP server on the wireless controller. We use the DHCP server on the wireless controller to give all AP's and management workstations IP addresses. In case this was a wireless controller DHCP broadcast issue we decided to give our management workstation a static IP address and we were able to establish contact with the controller via the web interface and where able to verify the controller was unable to manage the wireless AP's.
Actually, it would connect to the AP's, apply the configuration and then the AP's would enter into this "Not Connected State" in red writing. We would notice two or three AP's throughout the network would come online (via their uptime) for a few seconds and then die, returning to a 'Not Connected" state. We confirmed that the port configuration settings where not inconsistent from the original on the GSM7224P. We were worried that where was some sort of “binding” propriety issue with the WC7600v2 and non-Netgear equipment so we decided to use the Ethernet port on the M4100-12GF (which connected originally to the GSM7224P) and plugin into the WC7600v2 into that and we had the same affect. The wireless controller was unable to connect to the AP's, in fact we were no longer able to communicate to the wireless controller on that port on the M4100-12GF (we confirmed the port settings / VLAN settings where consistent).
After 5 hours of checking and rechecking configs and even thinking spanning tree was the problem (we set the 5120E to be the root primary bridge) we decided take everything apart and back step to the previous configuration. Before we undertook something so drastic we decided to bring the GSM7224P back into the mix and attach switch back to the into the same port as it was setup before on M4100-12GF and then connect the WC7600v2 into the same port it was originally connected on the GSM7224P.
Wouldn't you know it all worked within about 10 to 20 seconds it instantly found all the AP's and the wireless network started to function as normal. So we now have the following setup,
1 x 5120EI 24 port POE+
This switch performs all layer three routing and provides connectivity for all devices in the Core rack. It connects to the M4100-12GF via a single Ethernet link.
1 x M4100-12GF ProSafe 12-port Gigabit Fiber L2+ Managed Switch with PoE+
The M4100-12G serves as the switch that aggregates all the backbone runs throughout the site at the centre of the star / network and is connected to the 5120EI a via single Ethernet link and the GSM7224P via single Ethernet link.
1 x GSM7224P ProSafe 24-port Gigabit L2+ Managed Switch with PoE+
This switch provided Ethernet connectivity to the WC7600v2 has a single Ethernet connection the M4100-12GF ProSafe 12-port.
Now that everything was working, we decided to perform some testing.
The management workstation was connected to the 5120EI and was able to ping the WC7600v2 when it was connected to GSM7224P successfully. While we were pinging the device we quickly reconnected the WC7600V2 to the H3C/HP 5120EI and the device responded, it would ping for about 10 to 15 pings and then all of a sudden then die out and become inaccessible. Reconnecting the WC7600v2 to the GSM7224P the device responds instantly to ping for prolonged periods, move it to the 5120EI, pings for about 10 to 15 seconds and then drops off and vanishes.
We check the diagnostic logs of the 5120EI and we find nothing out of the ordinary. No error messages, nothing, plug the WC7600v2 to the M4100-12GF, even with the same VLAN's and port settings as the GSM7224P still fails but as long as you plug in the WC7600v2 into the GSM7224P everything works as expected.
Everything in my / team's gut tells us that it is the controller that is behaving badly and seems to be bound to this one switch. Can I please get a confirmation on this from Netgear if their some sort of security binding of the controller to a particular device? Or even get in contact with a local support member / team to confirm this strange behaviour.
Also the controller has no problem contacting 2 of the Netgear AP's that are running from the 5120EI switch. The 5120EI give us no error messages other than a “detected topology change via STP” and that it, the WC7600v2 pings for about 10 to 15 seconds.
We even tried to SSH into the wireless controller to monitor its status messages but from our understand command line access to the controller is only reserved to NETGEAR support personnel. We did connect a console cable and monitor as many messages as we could from the limited unauthenticated output the controller gave us. No help.
I suspect you will request that we provide you with the configs for all devices in question, which I am willing to do so.
I hope that if you made it this far you will be able to help.
PS We even disabled spanning tree on the Ethernet port on the 5120EI we used to connect the WC7600v2 just in case that was the issue. It could also be a backbone issue but we check those as well and the rest of the switches configures for missmatched VLAN's and the like.
Re: WC7600v2 only connects to one switch / strange behaviour
Hi John,
Thanks for the reply.
After some more of the gnashing of teeth, I decided to consult the manual of the 7600v2 once more and I realised that the ports on the 7600v2 are already set to TRUNK mode if you DON'T tick the Untagged management option in the web interface settings. Meaning management traffic the 7600v2 by default is TAGGED and the device is expecting the traffic to appear as such on the port.
Since I was following the configuration that was implement previously at the site I replicated the port config on 5120EI that was on the port the GSM7224 that originally had connect the 7600v2. The port on the 5120EI was configured as follows,
interface Gigabit 1/0/3 description ** Link to Netgear 7600v2 ** undo port trunk permit vlan 1 port link-type trunk port trunk permit vlan 200 400 to 410 port trunk pvid vlan 400 broadcast-suppression 20
Now knowing that VLAN 400 was the management VLAN for management traffic for the wireless network you would expect that there would be no problem in port 1/0/3 providing management traffic to the 7600v2 but alas, because of the PVID 400 statement the traffic would appear on the port as UNTAGGED traffic to the wireless controller and thus inaccessible.
On the GSM7224 the PVID port config allows for ANY type of 400 traffic both tagged and untagged by nature of the software on Netgear devices and it looks like the H3C 5120EI adheres to a stricter standard. Once I removed the PVID on port 1/0/3 statement, the 7600v2 performed as expected and configured.
I just find it so strange that the interpretation of such basic standards are so different between vendors.
Let us not even get started on Netgear's GSM7224 10.0.2.26 firmware igmp-snooping implementation that breaks cisco router traffic, but that will be a post for another time after some testing.
Re: WC7600v2 only connects to one switch / strange behaviour
Hi PeterA23,
It seems that you pretty isolated the issue already. Netgear wireless controllers doesn't have any propriety issues with other manufacturers. In order to assist you further on this issue, please open a chat / case online on NETGEAR Support and let them know the status of your network.
Re: WC7600v2 only connects to one switch / strange behaviour
Hi John,
Thanks for the reply.
After some more of the gnashing of teeth, I decided to consult the manual of the 7600v2 once more and I realised that the ports on the 7600v2 are already set to TRUNK mode if you DON'T tick the Untagged management option in the web interface settings. Meaning management traffic the 7600v2 by default is TAGGED and the device is expecting the traffic to appear as such on the port.
Since I was following the configuration that was implement previously at the site I replicated the port config on 5120EI that was on the port the GSM7224 that originally had connect the 7600v2. The port on the 5120EI was configured as follows,
interface Gigabit 1/0/3 description ** Link to Netgear 7600v2 ** undo port trunk permit vlan 1 port link-type trunk port trunk permit vlan 200 400 to 410 port trunk pvid vlan 400 broadcast-suppression 20
Now knowing that VLAN 400 was the management VLAN for management traffic for the wireless network you would expect that there would be no problem in port 1/0/3 providing management traffic to the 7600v2 but alas, because of the PVID 400 statement the traffic would appear on the port as UNTAGGED traffic to the wireless controller and thus inaccessible.
On the GSM7224 the PVID port config allows for ANY type of 400 traffic both tagged and untagged by nature of the software on Netgear devices and it looks like the H3C 5120EI adheres to a stricter standard. Once I removed the PVID on port 1/0/3 statement, the 7600v2 performed as expected and configured.
I just find it so strange that the interpretation of such basic standards are so different between vendors.
Let us not even get started on Netgear's GSM7224 10.0.2.26 firmware igmp-snooping implementation that breaks cisco router traffic, but that will be a post for another time after some testing.
