We can skip a error in GTID based replication by following steps:
STOP SLAVE;
set GTID_NEXT='SERVER_UUID:LAST_TRANSACTION_NUMBER+1' ;
BEGIN; COMMIT; SET GTID_NEXT="AUTOMATIC";
START SLAVE;
But if a replication is running with channel information, than how to skip the transaction for a particular channel ?
We can give "for channel" keyword in stop slave and start slave. But how to skip transaction for a particular channel, like in set GTID_NEXT command or what ?
In a replication topology GTID is a global unique identified for any transaction, therefore if transaction is required to skip, specifying channel becomes irrelevant here. It is similar to MySQL Replication Filters [MySQL 5.7] are global , or in other words, will be applied for all running replication channels