グリアンドロ

Sveta, Let me comment inline. >SHOW REPLICA STATUS exists since 8.0.22, therefore check regarding 8.1 is wrong. The ${mysql_version} variable pt-stalk uses to check versions only has major.minor values, so...

Sveta, New test case created as t/pt-stalk/pt-stalk-replication.t. Please review it and let me know what you think. ``` shell> prove -vw --trap --timer t/pt-stalk/pt-stalk-replication.t [22:05:50] t/pt-stalk/pt-stalk-replication.t .. ok 1 -...

Actually, it's not working even with curl now...

Wrong link used: https://github.com/maxbube/mydumper/releases/download/v0.12.5-3-zstd/mydumper-0.12.5-3-zstd.el7.x86_64.rpm Currently working link: https://github.com/mydumper/mydumper/releases/download/v0.12.5-3/mydumper-0.12.5-3-zstd.el7.x86_64.rpm

Hi Shlomi! Thanks for the quick turnaround on this :) I have now tested using `DelayMasterPromotionIfSQLThreadNotUpToDate=true` and `FailMasterPromotionIfSQLThreadNotUpToDate=false` and got the same results, in which the not up-to-date slave was...

Shlomi, >With async replication, how do we know that the replica is not up to date? We cannot guarantee consistency no matter what. Possibly I'm not understanding the question. Maybe...

Shlomi, I'll circle these back to the team, to gather more input. In the meantime, let me add some quick comments: Regarding point **4**: in these cases, lag reported is...