x670V-48x native stacking stack port utilization issue with multicast

  • 7 February 2018
  • 5 replies


When using 3 x670_48x switches in a stack (native) at a client, I'm experiencing high stack-port utilization. This reaches 100 percent on certain stack ports. I am using native v160 stacking.
The firmware used on these switches is, the license is CORE.

Setup explanation
The setup is used to stream multicast raw video within a VLAN to media devices (monitors and servers). The bandwidth of the stream is about 3,6 gb/s (uncompressed).
When contacting our supplier we were told that all multicast traffic will flow through the stack ports. This will cause packet drops on the stack ports and will saturate the stack ports.

I was under the impression that even with stacking, the multicast should stay within its own VLAN?

To resolve this issue, is it necessary to remove native stacking from our setup? If not, is there an alternative solution to our problem?

If there is any more information required to answer my question(s), feel free to post below.

Thanks in advance,

Erwin van Hoof

5 replies

Userlevel 7
Hello Erwin,

Is this a new setup? What command are you using to look at the utilization?
Userlevel 7
I've made this conversation public so others can respond. When it's set to private, only employees can see it.
Hello Patrick,

The setup is not a new setup, it has been running for a year now.
The command I use for the utilization:
show ports stack-ports utilization [/code]

Userlevel 7
Hello Erwin,

I found an issue we saw on 15.3 related to stack port utilization being high. It may or may not have been fixed in 15.4. However, 15.4 is currently unsupported code. I would recommend upgrading to patch 1-5. If the issue persist please open a case with GTAC for further review.
Thank you for doing the research to conclude there might be a firmware issue. We will try to update the firmware asap and see if the situation improves.