NSX Manager reports alarms for Edge Node "transmit ring buffer has overflowed" with an overflow percentage lower than 0.1%
book
Article ID: 336792
calendar_today
Updated On:
Products
VMware NSX Networking
Issue/Introduction
Symptoms:
NSX-T version is 3.0.0
NSX Manager reports alarms for Edge Node "transmit ring buffer has overflowed", similar to:
Edge NIC fp-eth0 transmit ring buffer has overflowed by 0.066608% on Edge node de51d19a-XXXX for over 60 seconds.
Edge Node logs (syslog.log) display message(s) similar to:
<178>1 2020-07-28T10:58:46.811Z ptk-edge-tn-01 NSX 4306 - [nsx@6876 comp="nsx-edge s2comp="nsx-monitoring" entId="79455a6a-8172-4ef0-8d42-ffd7ff88ed3c" tid="4396" level="FATAL" eventState="On" eventFeatureName="edge_health" eventSev="critical" eventType="edge_nic_out_of_receive_buffer"] Edge NIC fp-eth1 receive ring buffer has overflowed by 0.066608% on Edge node de51d19a-XXXX for over 60 seconds.
The overflow percentage reported is lower than 0.1%
Environment
VMware NSX-T Data Center VMware NSX-T Data Center 3.x
Cause
In NSX-T 3.0.0 the alarm trigger was designed with a low threshold of 0.01% and a short sample period of 60 seconds. If micro burst of traffic are seen this could result in false positive alarms.
Resolution
The issue is resolved in NSX-T 3.1.0, available at VMware Downloads. The threshold for the alarms has been changed to 0.1% and the sample period increased to 120 seconds.
Workaround: Alarms reporting an overflow percentage lower than 0.1% can be ignored.