Re: Replication breaks

[James Day]

1. This usually happens when a human gives a relay log position instead of an IO thread position to a CHANGE MASTER statement. that's diagnosed by looking at the end of the error log file on the slave and comparing the value at the last shutdown with the value being set.

2. You can generally fix this by using the correct position in a new CHANGE MASTER statement.

Sometimes it happens after there has been a RESET MASTER ont the master, erasing the required binary log files, or when they have been moved before replication hs finished with them.

The usual trick here is to examine the current settings with SHOW MASTER STATUS and SHOW SLAVE STATUS and compare to what is in the error logs, looking for inconsistencies.

James Day
Support engineer, MySQL AB