剑指offer 一些排序查找算法 leetcode等等 持续不断的更新
资源调优： 资源池分档： 分实时流和批处理
档位： 实时流：1C512M、1C512M、2C512M、3c512m, 1c1g、3c1g、1c2g、2c2g、3c2g 2c4g 6c4g 3c3g 3c6g 4c8g 8c8g 12c8g 8c16g 12c16g 12c24g 18c36g 24c48g
1c2g: executors:1 cores:1 memory:2048 spark.executor.memoryOverhead:512 spark.yarn.executor.memoryOverhead:512 spark.sql.shuffle.partitions:12 spark.default.parallesism:12 spark.buffer.pageSize:4M maxFailedTimes:5 maxWaitingBatch:10
2c2g: executors:2 cores:2 memory:1024 spark.executor.memoryOverhead:1024 spark.yarn.executor.memoryOverhead:1024 spark.sql.shuffle.partitions:36 spark.default.parallesism:36 maxFailedTimes:5 maxWaitingBatch:10
批处理：1c512m 2c1g 2c2g 2c3g 3c3g 4c3g 6c3g 2c4g 6c4g 3c6g 4c6g 9c6g 4c8g 12c9g 6c12g 9c12g 9c18g 18c18g 18c36g 15c9g 18c12g 4c48g 24c18g 50c60g
1c512m: executors:1 cores:1 memory:512 driver.memory.mb:1024 spark.executor.memoryOverhead:512 spark.yarn.executor.memoryOverhead:512 spark.sql.shuffle.partitions:12 spark.default.parallesism:12 spark.buffer.pageSize:4M
9c18g: executors:3 cores:3 memory:6144 driver.memory.mb:1536 spark.executor.memoryOverhead:1024 spark.yarn.executor.memoryOverhead:1024 spark.sql.shuffle.partitions:48 spark.default.parallesism:48
18c18g: executors:3 cores:6 memory:6144 driver.memory.mb:1536 spark.executor.memoryOverhead:2048 spark.yarn.executor.memoryOverhead:2048 spark.sql.shuffle.partitions:150 spark.default.parallesism:150