Maxscale: Client stops connecting after a while.
MaxScale 2.4.3
2 backend servers, one master, one slave.
Our transaction manager uses maxscale to connect to the master for writes and the slave for reads through readwritesplit.
After a while the transaction server stops connecting with a "|Error| [SQL] Error: can't connect to db"
The number of sessions arn't high.
When we connect the transaction server directly to the master server there is no problem.
Current config below :
[maxscale] threads=auto
[db1] type=server address=db1 port=3306 protocol=MariaDBBackend
[db2] type=server address=db2 port=3306 protocol=MariaDBBackend
[MariaDB-Monitor] type=monitor module=mariadbmon servers=db1,db2 user=util_user password=* monitor_interval=1000 auto_failover=true auto_rejoin=true
[MaxAdmin-Listener] type=listener service=MaxAdmin-Service protocol=maxscaled socket=default
[MaxAdmin-Service] type=service router=cli
[Read-Write-Listener] type=listener service=Read-Write-Service protocol=MariaDBClient port=3306 address=0.0.0.0
[Read-Write-Service] type=service router=readwritesplit servers=db1,db2 user=util_user password=** filters=HiFi enable_root_user=1 connection_keepalive=30s
[HiFi] type=filter module=hintfilter
[QLAFilter] type=filter module=qlafilter filebase=/var/log/maxscale/qla.log match=/.*/ flush=true log_data=service, session, date, user, reply_time, query log_type=unified newline_replacement=" LANG "
Answer Answered by Markus Mäkelä in this comment.
Please open a new issue on the MariaDB Jira under the MaxScale project. This way we have a more convenient place for logs and configuration files and we'll be able to track the issue.