Repeated multi-threaded slave replication failures

[monty solomon]

I enabled multi-threaded slave replication and we are experiencing repeated replication failures where the slave stops with an error message similar to this one.

                   Last_Errno: 1032
                   Last_Error: Coordinator stopped because there were error(s) in the worker(s). The most recent failure being: Worker 2 failed executing transaction 'c7258e28-4c06-11e7-b189-0a26f8645d98:242101388' at master log bin.000406, end_log_pos 906634781. See error log and/or performance_schema.replication_applier_status_by_worker table for more details about this failure or others, if any.

I executed stop slave; start slave; and it failed again after a while

                   Last_Errno: 1032
                   Last_Error: Coordinator stopped because there were error(s) in the worker(s). The most recent failure being: Worker 3 failed executing transaction 'c7258e28-4c06-11e7-b189-0a26f8645d98:242473841' at master log bin.000407, end_log_pos 436498711. See error log and/or performance_schema.replication_applier_status_by_worker table for more details about this failure or others, if any.

Sometimes I am unable to resume replication with the server reporting the error

                Last_IO_Errno: 1236
                Last_IO_Error: Got fatal error 1236 from master when reading data from binary log: 'The slave is connecting using CHANGE MASTER TO MASTER_AUTO_POSITION = 1, but the master has purged binary logs containing GTIDs that the slave requires.'

[monty solomon (monty+launchpad)]

Server version: 5.7.18-15-log Percona Server (GPL), Release 15, Revision bff2cd9

CentOS release 6.8 (Final)

[monty solomon (monty+launchpad)]

I created upstream https://bugs.mysql.com/bug.php?id=86641

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-1099