I've setup a Group Replication with two servers and I have a third that will be joining the cluster shortly. I have it setup with a single primary, so right now there is one writer and one reader. Both systems are the exact same hardware (they were ordered together) ... the primary is keeping up with the load with no problem. However the reader is falling more and more behind ... the difference between the last received GTID and the last executed GTID is growing at a very large rate .... in a few hours it has gone over 7 million and it's still growing. According to the replication_group_member_stats the reader has 0 transactions in the queue ... but the last executed vs the last received is showing a large backup. I'm not sure where to even begin to troubleshoot this. Does anyone have any suggestions?
↧