Seeing this error message "NX-SENDTO-Unable to send message of 64 bytes" in the domain log and how to prevent them
search cancel

Seeing this error message "NX-SENDTO-Unable to send message of 64 bytes" in the domain log and how to prevent them

book

Article ID: 332028

calendar_today

Updated On:

Products

VMware Smart Assurance Network Observability

Issue/Introduction

Symptoms:

Seeing a large number of messages similar to the example below in the AM/PM log.

NX-SENDTO-Unable to send message of 64 bytes to destination [<IP V6 Address>].; in file "/work/tancurrent/DMT-<VERSION>/116/smarts/clsapi/network/lib/NX_Socket.c" at line 243
PING-PINGER_EICMPECHO-Failure when sending icmp echo request.; in file "/work/tancurrent/DMT-<VERSION>/116/smarts/icmp-accessor/pinger/ICMP_Poller.c" at line 154

Environment

All Supported Smarts releases

Cause

The issue here is with machine IPv6 address configuration. The machine is not configured as Dual stack.
This error also occurs when machine where IP domain is running has IPv6 stack configured but no IPv6 address configured and IP domain is discovering dual stack network devices which has both ipv4 and ipv6 configured using ipv4 addresses.

Since the machine has no ipv6 address configured and ipv6 stack is configured, Smarts goes ahead and discovers IPv6 addresses using IPv4 address as discovery agent. Now when it comes to monitoring / polling it tries to poll those IPv6 address which it will fail.  Because the machine does not have ipv6 configured but IPv6 address is in the topology.  This is why the error messages will be printed every polling interval.

Resolution

Below are the two possible workarounds to avoid these errors:

  • Configure IPv6 address for the machine where IP domain is running.
  • Removing IPv6 stack from the OS may also work which will prevent discovery of IPv6 address.

After you make the changes on the machine, restart the IP domain inorder to take configuration changes into effect.

If you choose to ignore those errors it will not cause any issues with IP domain in the long run.
Only periodic error messages and logs will be rolled over.