We have upgraded 2 servers of our 4-server-MariaDB-cluster from 10.3.14 to 10.3.15 today and since then we had several slave-stops only on the 10.3.15-nodes because a table could not be found, but the table existed. It was always the same table and a "start slave;" solved the problem - there was no need to skip the query.

Last_Error: Error 'Table 'db1.table1' doesn't exist' on query. Default database: 'db1'. Query: 'UPDATE `table1` AS `cpm` INNER JOIN `table2` AS `cp` ON (`cpm`.`id` = `cp`.`id`) SET `cpm`.`flags` = CONCAT_WS(',', IF(`cpm`.`flags` = '', NULL, `cpm`.`flags`), 'DELETED') WHERE `cp`.`pid` = 1234 AND (`cpm`.`flags` IS NULL || FIND_IN_SET('DELETED', `cpm`.`flags`) = 0)'