Buffer Threshold Exceeded log

[networknub]

hello

I am currently using version 9.3(7) of n9k.
I have checked the logs and the following log occurs once a day.

%TAHUSD-SLOT1-4-BUFFER_THRESHOLD_EXCEEDED: Module 1 Instance 1 Pool-group buffer 90 percent threshold is exceeded!

What can I do to resolve that issue?

[marce1000]

              - FYI : https://bst.cloudapps.cisco.com/bugsearch/bug/CSCvz97156

 M.

[networknub]

I am using "N9K-X97160YC-EX" and this log occurs whenever the threshold reaches 90%. I don't think it's a bug.

[lanman1960]

Having this issue since November across multiple N9ks running 9.3(9).    What is weird is this problem had not shown up after our upgrades back in January 2023........the errors only showed up when client Juniper equipment was upgraded in this past November.      Client has had unstable BFD instances where the session would disconnect and reconnect intermittantly.     Not sure any of this is related but seems very coincidental.   

[it-devices]

Hello，Which version did you upgrade to?

[lanman1960]

Hi     The version we are currently running is 9.3(9).    Like I mentioned earlier in this post......these issues did not appear until the Juniper switches were upgraded that these 9000s are attached to.   The issue is still happening, and the client is waiting for Juniper to assist further.   We had other SMEs look at the Cisco configuration and they find no issues.   

[networknub]

What is your interface bandwidth, and have you tried increasing it?

[lanman1960]

The interfaces between the Cisco gear and Junipers are 4) 100 gig links and no we have not tried to add additional links. Those links are in port channels connected spanning across the 2 9000s.

[networknub]

In my case, I solved the problem by moving the offending interface to a different interface (called an instance) with separate buffers.

You may also find the article below helpful

https://www.cisco.com/c/en/us/support/docs/switches/nexus-9000-series-switches/217340-understand-the-tahusd-buffer-threshold-e.html

[it-devices]

Hello, thank you for your response. I have encountered this issue on both 9.3.(9) and 9.3.(12). Our current network architecture involves connecting 93240 to 9504, with a 100G*4 Ethernet channel in between. The issue is still under investigation, and if a solution is found, I will share our resolution.

[lanman1960]

What is weird this infrastructure is VxLAN made up of 24) 9000s in a Leaf, Spine and TOR deployment. We were running 9.3(7) and upgraded to 9.3(9) in November/December of 2022 and never saw this issue before. The only time we saw this is after the Juniper switches were upgraded this past November.
The Junipers only connect to the 2) Bdr-Leafs in a 4)100G port channel. WE are now seeing the error on both the Border Leafs and across 10 TORs (vPC pairs ) so multiple interfaces on multiple switches.
Thanks I appreciate it.



Disclaimer: This email and any attachments are sent in strictest confidence for the sole use of the addressee and may contain legally privileged, confidential, and proprietary data. If you are not the intended recipient, please advise the sender by replying promptly to this email and then delete and destroy this email and any attachments without any further use, copying or forwarding.

[lanman1960]

Hello Any update if a resolution has been found?