The resolution for this issue was included in NSX-V version 6.4.6. If a customer is running a version below 6.4.6 the recommendation is to upgrade.
Workaround:
As a workaround, a log filter can be added to prevent this log from flooding the log file:
- Log in to the ESXi Shell as root.
- In the /etc/vmware/logfilters file, add the following entry to create a log filter.
numLogs | ident | logRegexp
where:
For example, to set a limit of maximum two-log entries from the hostd component for messages that resemble the SOCKET connect failed, error 2: No such file or directory phrase with any error number, add the following entry:
2 | hostd | SOCKET connect failed, error .*: No such file or directory |
- numLogs sets the maximum number of log entries for the specified log messages. After reaching this number, the specified log messages are filtered and ignored. Use 0 to filter and ignore all the specified log messages.
- ident specifies one or more system components to apply the filter to the log messages that these components generate. For information about the system components that generate log messages, see the values of the idents parameters in the syslog configuration files. The files are located in the /etc/vmsyslog.conf.d directory. Use a comma-separated list to apply a filter to more than one system component. Use * to apply a filter to all system components.
- logRegexp specifies a case-sensitive phrase with Python regular expression syntax to filter the log messages by their content.
- In the /etc/vmsyslog.conf file, add the following entry to enable the log filtering capability.
- Run the esxcli system syslog reload command to reload the syslog daemon and apply the configuration changes.
This workaround can also be found here:
https://docs.vmware.com/en/VMware-vSphere/6.7/com.vmware.vsphere.monitoring.doc/GUID-D0D77526-65DC-4D08-A52F-51D5B0DAF8C3.html