Without going into too much detail, we have a fairly simple setup with a fairly confusing issue. We are running a uBR10k to a Nexus 3548 and that to an APEX3k for our M-CMTS solution. We have 8x 3G60s running via fiber to the Nexus, then a single 10G fiber from the Nexus to the APEX3k. Every now and then, we actually drop our DEPI tunnels. Random tunnels, random times, and no help from the logs.
UBR Logs (nothing about the initial down):
3236336: Jun 9 15:58:14.828: %DEPI-6-SESSION_UP: Depi session Modular-Cable5/0/2:15, changed state to up
3236339: Jun 9 15:58:21.752: %DEPI-6-SESSION_UP: Depi session Modular-Cable5/0/2:23, changed state to up
3236340: Jun 9 15:58:29.672: %DEPI-6-SESSION_UP: Depi session Modular-Cable6/1/1:14, changed state to up
3236345: Jun 9 15:58:45.856: %DEPI-6-SESSION_UP: Depi session Modular-Cable6/0/1:3, changed state to up
Nexus Logs:
Nothing....
APEX Logs:
Mon, Jun 9, 2014 15:58:10 Received StopCCNMsg; result code: 2; error code 6; Too many retransmits to 192.168.250.5
192.168.250.5 is the IP on the APEX interface.
Anybody have any clue? Much more information can be provided if needed.
Thanks,
Everett
Hi,
We are not using APEX3k but, I had a talk with some Arris sales engineer and he told me that the APEX3k has some L2TP issue.
The issue was that the APEX e-qam was randomaly was not respons to the hello messages that thew CMTS sends. Because of that the CMTS assumes that the depi session is down and restart the session.
Without getting into to much detais it looks like this is the problem you're suffering from.
As far as I understood it was solved by a SW version in the APEX side.
You need to contact the APEX3k support team.
Best regards,
Asaf
Hi Everetts,
I'm sorry I can't help you with your problem. But I'm very interested about how APEX3k have worked with your CMTS10k. Are you doing tests or this is your actual production environment?
Why did you choose APEX3K instead of RFGW1 or RFGW10 for your modular solution?
I appreciate your help. I hope someone could help you with your problem.
We choose to go with the Apex3k because our future needs to run VoD services through it, as a CCAP-type setup. We had considered the RFGW10; however, the APEX3k was a better and cheaper fit. This is our production environment, and the APEX3k works well with the uBR10k. We did aggregate all the 1G links from the 3G60s, into a single 10G link via a Nexus switch, and everything seems to be a-okay. We have had this all in production for almost 5 months now.
you said the Nexus logs are empty. Have you checked for errors on the individual ports of the switch?
Any interface errors reported by your switch?