I have two VM machines (with exact hardware-software configuration on two ESXI Servers in same rack and same subnet too) that is running with nginx + PHP-FPM + and MariaDB 10.4.
in each ESXI server I have just 1 VM with 24G RAM, 24 Core CPU
On standalone nodes the TPS is 8000 avg.
But when I switched to galera Master-Master replication mode with below conf, The TPS is decrease into 1100 avg!
each server's app using It's local DB.
here is my.cnf
[mysqld]
skip-name-resolve
binlog_format=ROW
wsrep_on=ON
wsrep_replicate_myisam=ON
wsrep_provider=/usr/lib64/galera-4/libgalera_smm.so
wsrep_cluster_address="gcomm://192.168.7.11,192.168.7.12"
wsrep_cluster_name="my_cluster"
wsrep_node_address="192.168.7.11"
wsrep_node_name="node1"
wsrep_sst_method=rsync
wsrep_slave_threads=16
innodb_flush_log_at_trx_commit=2
innodb_flush_method = O_DIRECT
What am I missing?
