A Lord from test environment:
The primary and secondary libraries are set parameters after sync_binlog=1000, mandatory master database server, when the power is from the library slave_io complains 1236 mistakes, and may be more than the main library data from library data (for the library of the binlog cache refresh to binlog disk files, but reached the from the library implemented),
Don't consider data consistency, after restart the master database server, from the traditional library in order to change the master to master_log_file='mysql - bin. XXXXX', master_log_pos=XXXX; Can the new pull slave_io threads continue to duplicate the main library data,
question: but in the case of gtid, can adjust (narrow) from library gtid size consistent with the main library? In the case of a fixed main library from the library how to pull up slave_io threads continue to copy the main library data?
Thank you very much!
The main library status:
From the library status:
