Connectivity Issues

[netguyz08]

Just deployed Thunderbolt in my office network today, and suddenly noticed all of the Logmein icons in the System Trays on the servers and PCs had a yellow exclamation point, meaning there is a network error and they can't connect out.

Connected to a client through Logmein and kept getting kicked out during the session. The session would disconnect and I'd have to reconnect. Unplugged the Thunderbolt device only (still have the ESW 520 8 port and Cisco WAP4410N running) and Logmein is solid again. All of the icons on the computers are back to normal, too.

[davebainum]

Hi all,

I'm going to chime in on this thread as well, because we had seen some strange connectivity issues when trying to use the Cisco switches as well.  It's possible that we just didn't have the Cisco switches properly configured, I'm still not 100% sure, but here's what happened:

- When we would attempt to use the Cisco switch as the "core switch" (e.g. the main switch for the installation site, directly plugged into the TBA adaptor, and also plugged into the cable modem etc.), we would have difficulty then daisy-chaining other switches throughout the network.  While devices were plugged directly into the Cisco switch, it would generally work, but other switches in turn plugged into that main Cisco switch could not get DHCP assignments or access other resources on the core Cisco switch.  However they could get out to the Internet if they had a static assignment.

- In another case at another site, when we plugged in the TBA directly to the core switch, it never was able to check into the TBA at all.  The moment we removed the Cisco switch as the "core switch", and replaced it with a Linksys/Cisco WRT54G which also acted as an inexpensive router/firewall/wireless point, TBA checked in fine when connected to a different PoE Switch that was daisy-chained off of that WRT54G

- These are all kind of strange problems, but mysteriously all of these problems would go away when we used an alternate switch (eg. Netgear) as the "core switch" - TBA would work properly.

- When we've used the Cisco switches as secondary switches (not the "core switch" plugged directly into the TBA), so far they have not seemed to have issues, however we're going to continue to do further testing.

I will be curious to see if there are more insights into whether this is just a switch misconfiguration issue on our part (??) or whether there's something else that's more a Thunderbolt particular thing going on.

Regards,

-- Dave Bainum (dbainum@ritetech.net),   PMP*
tinyurl.com/plf8wz>
RiteTech LLC / www.ritetech.net / +1 (703) 561-0607
Creators of the www.RPConnect.net suite of applications

[john.grange]

Hi Dave,

I too have witness situations similar to this, one of the problems was the way the switch was setup by default to do CDP / Voice Vlan. But I am still suspecting that there might be an issue with the switch. Perhaps I will try your solution and eliminate the switch although I do not have another POE switch here to test with.

[john.grange]

I just removed the wireless device from the Cisco switch and plugged it into one of my other dumb switches with the power adaptor and now I am able to get wireless from all the devices at this location, both N and G at once. Previously If I rebooted the device I could only get wireless on the first couple of devices.

[netguyz08]

Dave,

Just to give you my experience, I receivced 520 and 540 ESW 8 port PoE switches. Neither one gave me a different experience at my office with the TBA - the TBA's firmware chattering away on the network was definitely the issue for my network connectivity problems. At another site, the TBA worked fine in the 540 which was the secondary switch. The main switch there was a 10/100 NorTel BayStack 24 port.

[john.grange]

The switch here is definitely causing issues at least with the wireless, I have now had the wireless on a separate switch all day and it has been completely reliable. My only other thought is that it might have to do with POE vs self powered, I will investigate that further tomorrow.

[netguyz08]

John,

Let me know what you find. the WAP4410 has been realiable for me on the 520 ESW switch, using PoE. Curious if there will be a setting change in the switches we all need to look for. I can post my config if needed.

[john.grange]

Please do, I have made quite a few changes to the switch including trying with and without DHCP relay.

[Brian Bergin]

Has the LAN cable been changed out?  A bad cable could certainly cause the connectivity problems being described here.

[john.grange]

You know I am using the provided cable and never thought about that, I will give it a try.

[Marcos Hernandez]

Something just occurred to me to: make sure your Smartport role is "Access Point" on the port connecting to the WAP4410N. See if that helps.

I have separately emailed the Product Team for the ESW, asking them if they know of issues like this.

Marcos

[john.grange]

Yes it has been set to Access Point since day one. Also we are going on 10+ hours now of stability, I do not think the problem is actually related to the Access Point any more at all, but leaning towards the switch.

[netguyz08]

John,

Here are the details on my ESW 520 10/100 PoE switch that the WAP4410N has been stable on. I haven't modified the config at all btw, and used an ethernet cable I had laying around. And it is on port 2, so I am just showing the config for that (from the telnet session into the switch)...

Product ID : ESW-520-8P

Versions
========

  SW Version:            2.1.16 (Date:  08-Sep-2009, Time:  15:59:03)

  Boot Version:          1.0.0.07 (Date:  08-Sep-2009, Time:  15:18:06)

  HW Version:            V01

System Mode
===========

Current System Mode:            Layer 2

System Mode:                    Standalone

Port Configuration
==================
                       Auto                Flow
  Port   Enable  Neg.  Spd/Dpx Ctrl
  ------------------------------------

e2       ENABLE   On   Auto    Off

PoE Configuration
=================
  Port   Priority     PoE
  ------------------------------

e2      LOW          ENABLE

Port Status
===========
                                              Flow
  Port   Enable   Link  Spd/Dpx Ctrl
  -----------------------------------

e2      ENABLE   UP    100F    Off

PoE Status
===========
                        Power         Power
  Port   Status   Allocation    Consumption
  ------------------------------------------

e2      ENABLE     15.4 W       4900 mW

There were some other screens, too. Let me know if you need me to look up anything else.

-Ed

[john.grange]

Finally figured out this mystery, it was the port security that was turned on by default. This is why the device would work flawlessly in my other switch, because it was plugged into the uplink port which has port security turned off. I can't believe I missed this setting.

[Marcos Hernandez]

I would have expected the Smartport role to take care of this. I will circle back with engineering to see if this setting might be affecting other situations.

Thanks a lot John!

Marcos

[Michael Holloway]

Great to hear that you sorted it out John!  Thanks for reporting back.

-mike