Two MySQL Connectors crashes Aurora MySQL 5.7

[C.AME]

Hello,

We were trying to run two new debezium MySQL connectors in our Kafka Connect environment. Initially we were only one (in snapshot mode). When we plugged two new connectors (fresh ones), in binlog mode, they took the MySQL (AWS RDS Aurora - 5.7) instance down.

Aurora logs:

210908 20:55:43 server_audit: Audit STARTED.
Found DAS config file, trying to load DAS switcher from DAS config file.
2021-09-08 20:55:43 47220272493120:[DAS][INFO]: Calculated persistence threads 4
aurora_enable_das:0
210908 20:55:43 server_audit: server_audit_incl_users set to ''.
210908 20:55:43 server_audit: server_audit_excl_users set to ''.
2021-09-08T20:55:43.137657Z 0 [Note] Recovering after a crash using /rdsdbdata/log/binlog/mysql-bin-changelog
2021-09-08T20:55:43.682702Z 0 [Note] Starting crash recovery...
2021-09-08T20:55:43.682758Z 0 [Note] Crash recovery finished.
2021-09-08T20:55:43.717719Z 0 [Note] InnoDB: Last MySQL binlog file position 0 5783487, file name mysql-bin-changelog.000277
2021-09-08 20:55:43 0x2af253917240 [PROGRESS_INDICATOR: in OSCAR_BINLOG_INIT_GTID] Reading from file /rdsdbdata/log/binlog/mysql-bin-changelog.000277 at position 2719744
2021-09-08T20:55:44.257217Z 0 [Note] Skipping generation of SSL certificates as options related to SSL are specified.
2021-09-08T20:55:44.257772Z 0 [Note] Skipping generation of RSA key pair as key files are present in data directory.
2021-09-08T20:55:44.257840Z 0 [Note] Server hostname (bind-address): '*'; port: 3306
2021-09-08T20:55:44.257870Z 0 [Note] IPv6 is available.
2021-09-08T20:55:44.257877Z 0 [Note] - '::' resolves to '::';
2021-09-08T20:55:44.258008Z 0 [Note] Server socket created on IP: '::'.
2021-09-08T20:55:44.258095Z 0 [ERROR] Another process with pid 3010 is using unix socket file.
2021-09-08T20:55:44.258099Z 0 [ERROR] Unable to setup unix socket lock file.
2021-09-08T20:55:44.258102Z 0 [ERROR] Aborting

...

...

2021-09-08T20:57:31.130830Z 452 [Note] Start binlog_dump to master_thread_id(452) slave_server(2), pos(mysql-bin-changelog.000277, 5783487)
2021-09-08T20:57:33.946373Z 453 [Note] Start binlog_dump to master_thread_id(453) slave_server(184058), pos(mysql-bin-changelog.000277, 5783487)
/etc/rds/oscar-start-cmd: line 39: 26479 Segmentation fault (core dumped) /rdsdbbin/oscar/bin/mysqld --core-file --innodb_shared_buffer_pool_uses_huge_pages='1' "$@"
2021-09-08T21:17:42.384973Z 908 [Note] InnoDB: Sync to disk of `schema`.`ONE_OF_THE_TABLES_OF_THE_NEW_CONNECTOR` started.
2021-09-08T21:17:42.384993Z 908 [Note] InnoDB: Stopping purge
2021-09-08T21:17:42.385052Z 908 [Note] InnoDB: Writing table metadata to './schema/ONE_OF_THE_TABLES_OF_THE_NEW_CONNECTOR.cfg'
21:17:42 UTC - mysqld got signal 11 ;
This could be because you hit a bug. It is also possible that this binary
or one of the libraries it was linked against is corrupt, improperly built,
or misconfigured. This error can also be caused by malfunctioning hardware.
Attempting to collect some information that could help diagnose the problem.
As this is a crash and something is definitely wrong, the information
collection process might fail.

Connector configuration:

{
    "connector.class": "io.debezium.connector.mysql.MySqlConnector",
    "transforms.unwrap.delete.handling.mode": "drop",
    "database.history.consumer.sasl.jaas.config": "CREDENTIALS",
    "database.history.kafka.topic": "dbhistory.tables.1",
    "transforms": "unwrap,Reroute_table1,Reroute_table2",
    "transforms.Reroute_table2.regex": "^dbname\\.schema\\.(table2)$",
    "value.converter.value.subject.name.strategy": "io.confluent.kafka.serializers.subject.TopicNameStrategy",
    "database.history.consumer.security.protocol": "SASL_SSL",
    "key.converter.key.subject.name.strategy": "io.confluent.kafka.serializers.subject.TopicNameStrategy",
    "transforms.unwrap.drop.tombstones": "true",
    "transforms.unwrap.type": "io.debezium.transforms.ExtractNewRecordState",
    "database.history.producer.sasl.jaas.config": "credentials",
    "database.history.producer.sasl.mechanism": "PLAIN",
    "database.user": "cdc-reader",
    "transforms.Reroute_table2.type": "org.apache.kafka.connect.transforms.RegexRouter",
    "database.server.id": "3",
    "database.history.producer.security.protocol": "SASL_SSL",
    "database.history.kafka.bootstrap.servers": "host",
    "database.server.name": "dbname",
    "transforms.unwrap.add.headers": "db",
    "database.port": "3306",
    "transforms.Reroute_table2.replacement": "cdc.dbname.table2",
    "database.hostname": "hostname",
    "database.password": "password",
    "name": "cdc-binlog-1",
    "database.history.store.only.monitored.tables.ddl": "true",
    "transforms.unwrap.add.fields": "table",
    "table.include.list": "schema\\.(table1|table2)",
    "transforms.Reroute_table1.replacement": "cdc.dbname.table1",
    "transforms.Reroute_table1.regex": "^dbname\\.schema\\.(table1)$",
    "database.history.consumer.sasl.mechanism": "PLAIN",
    "transforms.Reroute_table1.type": "org.apache.kafka.connect.transforms.RegexRouter",
    "database.include.list": "schema",
    "snapshot.mode": "schema_only"
}

We made sure that the server id were unique amongst the running connectors.

We dug into the binlog reader library without much success understanding why is this setup killing the database.

Does anyone has any idea? Any hint would be greatly appreciated.

Thank you very much!

Setup:

• MySQL 5.7.12 (AWS RDS Aurora)
• debezium 1.6
• Kafka Connect, single instance.
• The CDC connectors were running in standalone mode, with a single task
  

[C.AME]

Typo fix: "Initially we did have only one"