Bug #92250 replication slave cannot address positions in binlog file > 4G
Submitted: 30 Aug 2018 19:39 Modified: 31 Aug 2018 4:33
Reporter: Bill McCaffrey Email Updates:
Status: Duplicate Impact on me:
Category:MySQL Server: Replication Severity:S2 (Serious)
Version:all OS:Ubuntu (16.04)
Assigned to: CPU Architecture:Any

[30 Aug 2018 19:39] Bill McCaffrey
COM_BINLOG_DUMP binlog-pos value is only 32 bit so slave i/o thread start fails when the log position is greater than 4 gig.

How to repeat:
Load a large table using "load data infile" on the master, big enough to make the binlog greater than 4 gig. Break replication slave connection on the slave while position is greater than 4g and then try to restart replication.
I am seeing this problem while using row-based bin logging, I haven't tried statement based. 
The issue is also apparent in mysqlbinlog utility when passing a starting position greater than 4g.

Suggested fix:
break up bin logs into smaller chunks. This may require having a transaction be allowed to span multiple log files instead of being contained in a single binlog.
[31 Aug 2018 4:33] MySQL Verification Team
Hello Bill,

Thank you for the report.
This is duplicate of Shane's Bug #55231, please see Bug #55231