04-04-2018 10:33 AM - edited 03-20-2019 10:02 PM
I would like to request further investigation on this bug. We believe it to be causing packet drops at the endpoints. After we rebooted the host and the endpoints it seems to have fixed the issue.
05-18-2018 09:33 AM
Hi, we are seeing a packet loss issue on our Call manager v9.1.2.16900-2 running on VMware 6.0. We also see this "KERNEL assertion" error on the console of the CM subscribers.
This is happening at a site level. Phones at specific sites are not able to register to CM. These affected sites are fragmenting the REFER SIP message which is TCP. We constantly see the first part of the fragmented packet (60 bytes) being dropped.
When we switch the signaling to UDP it works or if we set the DF bit to 1 it works because that 1514 byte packet is not being fragmented.
We were able to take a packet captures at the physical NIC of the host and we still see the packet drop. I'm not sure if a reboot would fix our problem.
Can you please provide me with some insight with the issue that you experienced as well the type of packet loss that you saw and how it affected your environment.
Thanks in advance!
05-21-2018 01:46 PM
05-24-2018 07:00 AM
Thanks! I appreciate it. Here's what we have found:
We have determined that our 7k is receiving the packets from our WAN router and is dropping the packet on interface connecting to the UCS.
The specific packet is 60byte fragmented packet that is part of a SIP REFER message. This issue is only seen at sites that fragment the SIP REFER message.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide