NCS2006 MSTP - Constantly losing connection to EMS or Craft

[dmendesptd]

We see this issue quite often, we're staging an MSTP deployment of an NCS2006 M6 with dual TNC-Es, and we constantly lose connection to the shelf in CTC, making it difficult to setup some initial configuration parameters. This issue occurs on multiple shelves we have deployed, but not all of them. Some of the variables we've changed in order to rectify this issue:

-Used multiple laptops

-Used many different copper cables

-Tried different switch platforms

-Tried autonegotiate vs locked 100/full

-Tried both EMS and CRAFT

-Tried multishelf disabled and node controller modes

-Tried with and without a protection unit (dual TNC)

-Replaced TNCs, ECUs, and LCDs.

-Tried versions 10.6, 10.8, and 11.0

All of which never rectified this issue. We never lose packets to the shelf itself, our CTC cache just loses and regains communication with it. TL1 never disconnects either. Currently, I have this in our lab and I'm trying to find the secret to getting this to stop. I'm wondering if anyone else with experience with these MSTP shelves encountered the same type of issue and know of a fix? Fielding for a response here before engaging TAC.

Thanks!

[Ryan Gadwood · Accepted solution]

Hi,

It looks like you may have a firewall blocking the TCP callback from the node to the PC. You can test one of two things

• Temporarily disable the firewall on your laptop/PC and connect directly to the node.
• Enable Proxy on the node while you have access. Provisioning-->Network-->General. Enable Socks Proxy only

Thanks.

[dmendesptd]

Enabling the SOCKS Proxy yielded stability in CTC. Thank you so much!