>It look like the same issue as I reported in BUG #15151. >https://www.postgresql.org/message-id/flat/152353560977.31233.11406243320876395727%40wrigleys.postgresql.org >(unfortunately I didn't have time to explore it further then).
>You have good description of what's going on under hood.
Hi，I read the BUG #15151. You have say that it does not help to restart slave.
But in 15567, either 'restart slave' or 'move the wal segment from master to slave' can make the slave nomal.