This article addresses network stability issues that may occur due to high CPU utilization within environments where NSX and vRNI (or other CLI polling software) are used together for data collection and analysis.
Symptoms:
NSX for vSphere 6.3.5
vRealize Network Insight uses a CLI method to pull relevant Network Information from the Controllers. This procedure creates a new log file per action on the NSX Controller, up to one new log file per second. The logrotate process is unable to cope with the large amount of individual log files created.
Note: vRNI is just an example for a Software triggering this issue; other VMware or third party Software, or simple scripts used to login to the NSX Controllers CLI in short intervals, may trigger the same behavior.
This is resolved in NSX 6.3.6 and NSX 6.4.1
vRNI 3.7 uses a different mechanism to pull most information from the NSX Controllers, which reduces the amount of queries and therefore log files. A change in the logging mechanism for the CLI access will be included in an upcoming NSX Release.
Workaround:
Polling Information from the NSX Controllers can be disabled on vRNI with the following steps:
This will disable the collection of data from NSX Controller, however the data collection for other aspects of the environment will continue, including getting data from other components of the NSX Manager.
Impact/Risks:
Controllers might fail, leading to potential network disruption.