MySQL 8.x InnoDB Cluster with Router

[Sithivinayagam Nadar]

Hello Members,

Need your support and inputs to diagnose and resolve the problem with error ID: 2006: MySQL server has gone away

Setup & resources deatils

1.) InnoDB Cluster with three node with Ubuntu 20.04 & MySQL Community Server and MySQL Shell with version: 8.0.22-1ubuntu20.04
IP Details:
192.168.48.21
192.168.48.22
192.168.48.23

2.) this InnoDB cluster is running in Multi Primary Mode

3.) MySQL Router application with version: 8.0.22-1ubuntu20.04 is installed in a separate node along with sysbench version: 1.0.18+ds-1
IP Details:
192.168.48.29

4.) Router and InnoDB Cluster are connecting properly and able to do basic activities like DB, User creations, etc

Problem details

1.) We are observing this erro ID: 2006, while using sysbench to do data preparation by using the following command:

sysbench \
/usr/share/sysbench/oltp_common.lua \
--threads=1 \
--mysql-user=sbtest \
--mysql-db=sbtest \
--mysql-password='xxx@yyyy' \
--mysql-port=6446 \
--mysql-host=192.168.48.29 \
--tables=20 \
--table-size=100000 \
prepare

2.) But the same command was executed properly with the following command: when connected directly to one of the Node in this InnoDB Cluster

sysbench \
/usr/share/sysbench/oltp_common.lua \
--threads=1 \
--mysql-user=sbtest \
--mysql-db=sbtest \
--mysql-password='xxx@yyyy' \
--mysql-port=3306 \
--mysql-host=192.168.48.21 \
--tables=20 \
--table-size=100000 \
prepare


Here i am clueless with my limited knowledge of both MySQL InnoDB Cluster and MySQL Router

I could see some suggestions to use "max_allowed_packet" in the node configuration, but this doesn't help

Please share the suitable process / steps to troubleshoot this issue further.

Thanks / SG