Hello,
I have a v8.0.34 4 node cluster in Group Replication Single Primary. I have a MySQL Router on v8.0.35 routing WRITE traffic to the Primary DB.
I've been getting the errors below and the only solution that I have found thus far is to flush hosts on the Primary DB and restart the MySQL Router getting he errors.
I've been getting the following error on the router:
2023-10-11 12:02:22 routing INFO [7fab556f8640] Stop accepting connections for routing routing:group_rw listening on 3306
2023-10-11 12:02:22 routing ERROR [7fab54ef7640] no backend available to connect to
Sometimes the Router has timeout errors
On the Primary database, here is some information:
select * from performance_schema.host_cache\G
*************************** 1. row ***************************
IP: routerIP
HOST: router hostname
HOST_VALIDATED: YES
SUM_CONNECT_ERRORS: 1
COUNT_HANDSHAKE_ERRORS: 20
FIRST_SEEN: 2023-10-10 15:53:16
LAST_SEEN: 2023-10-11 12:02:14
FIRST_ERROR_SEEN: 2023-10-10 16:09:54
LAST_ERROR_SEEN: 2023-10-11 12:02:14
select * from information_schema.CONNECTION_CONTROL_FAILED_LOGIN_ATTEMPTS;
+---------------------------------+-----------------+
| USERHOST | FAILED_ATTEMPTS |
+---------------------------------+-----------------+
''@'issp-mr1.ia.local' | 20 |
+---------------------------------+-----------------+
This ''@'issp-mr1.ia.local' is causing the Router from connecting to the DB Server. SUM_CONNECT_ERRORS increments to 1 on the DB server and prevent future connections from the router to it.
There are NO errors on the DB /var/log/mysql/error.log
How can this be sent - ''@'issp-mr1.ia.local'? Blank username. What could be the problem here. There are "too many connection" errors. The MAX_CONNECTION_ERRORS and MAX_CONNECTIONS are set fairly high, both over 1000.
Faulty code, bug in DB or Router or MySQL Connector J? Is there any way I can mitigate or resolve?
This has been going on for over 2 months about every 2-3 days.
https://www.reddit.com/r/mysql/comments/16z73ex/mysql_8_router_randomly_cannot_connect_to_database/