开发者社区> 问答> 正文

发现flinksql写hive比写hdfs慢很多

两个job,都从同一个kafka读数据,一份写入hdfs,一份写入hive,都是分钟分区,并发都是200。运行一段时间后发现写hive要落后hdfs很多,而且hive任务对应的hdfs路径下,某一分区内的文件甚至跨度2个小时之久。大家遇到过这种情况没 附上对应ddl hive: CREATE EXTERNAL TABLE hive_table ( log_timestamp BIGINT, ip STRING, raw STRING ) PARTITIONED BY (day STRING, hour STRING,minute STRING) STORED AS PARQUET TBLPROPERTIES ( 'parquet.compression'='SNAPPY', 'sink.partition-commit.policy.kind' = 'success-file', 'sink.partition-commit.success-file.name' = '_SUCCESS' );

Hdfs:

CREATE TABLE hdfs_table ( log_timestamp BIGINT, ip STRING, raw STRING, day STRING, hour STRING,minute STRING ) PARTITIONED BY (day , hour ,minute) WITH ( 'connector'='filesystem', 'path'='hdfs://xx/test.db/hdfs_table', 'format'='parquet', 'parquet.compression'='SNAPPY', 'sink.partition-commit.policy.kind' = 'success-file’, 'sink.partition-commit.success-file.name' = '_SUCCESS' );

实际hdfs文件对比:

-rw-r--r-- 3 hadoop hadoop 1514862 2020-11-26 09:26 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-150-824 -rw-r--r-- 3 hadoop hadoop 10798011 2020-11-26 09:34 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-151-830 -rw-r--r-- 3 hadoop hadoop 4002618 2020-11-26 09:35 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-151-831 -rw-r--r-- 3 hadoop hadoop 8057522 2020-11-26 09:51 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-152-844 -rw-r--r-- 3 hadoop hadoop 6675744 2020-11-26 09:52 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-152-845 -rw-r--r-- 3 hadoop hadoop 4062571 2020-11-26 09:51 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-153-844 -rw-r--r-- 3 hadoop hadoop 10247973 2020-11-26 09:52 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-153-845 -rw-r--r-- 3 hadoop hadoop 483029 2020-11-26 09:53 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-153-846 -rw-r--r-- 3 hadoop hadoop 9440221 2020-11-26 09:16 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-154-816 -rw-r--r-- 3 hadoop hadoop 5346956 2020-11-26 09:17 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-154-817 -rw-r--r-- 3 hadoop hadoop 4940718 2020-11-26 09:51 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-155-844 -rw-r--r-- 3 hadoop hadoop 9687410 2020-11-26 09:52 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-155-845 -rw-r--r-- 3 hadoop hadoop 51998 2020-11-26 09:53 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-155-846 -rw-r--r-- 3 hadoop hadoop 3518 2020-11-26 09:37 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-156-833 -rw-r--r-- 3 hadoop hadoop 13801987 2020-11-26 09:39 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-156-834 -rw-r--r-- 3 hadoop hadoop 963288 2020-11-26 09:40 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-156-835 -rw-r--r-- 3 hadoop hadoop 6036601 2020-11-26 09:27 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-157-825 -rw-r--r-- 3 hadoop hadoop 8864235 2020-11-26 09:29 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-157-826 -rw-r--r-- 3 hadoop hadoop 10865872 2020-11-26 09:37 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-158-833 -rw-r--r-- 3 hadoop hadoop 4031077 2020-11-26 09:39 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-158-834 -rw-r--r-- 3 hadoop hadoop 228350 2020-11-26 09:09 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-159-811 -rw-r--r-- 3 hadoop hadoop 14661395 2020-11-26 09:11 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-159-812 -rw-r--r-- 3 hadoop hadoop 5451995 2020-11-26 09:29 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-160-826 -rw-r--r-- 3 hadoop hadoop 9149301 2020-11-26 09:30 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-160-827 -rw-r--r-- 3 hadoop hadoop 4731543 2020-11-26 09:30 hdfs://xxx/test.db/hive_table/day=2020-11-26/hour=08/minute=59/part-17bacc1b-cf96-463b-9cb7-aa318a05936c-161-827

-rw-r--r-- 3 hadoop hadoop 5950562 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-17-1288 -rw-r--r-- 3 hadoop hadoop 8922364 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-170-1287 -rw-r--r-- 3 hadoop hadoop 5898257 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-170-1288 -rw-r--r-- 3 hadoop hadoop 8848292 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-171-1287 -rw-r--r-- 3 hadoop hadoop 5893106 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-171-1288 -rw-r--r-- 3 hadoop hadoop 8905617 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-172-1287 -rw-r--r-- 3 hadoop hadoop 5800338 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-172-1288 -rw-r--r-- 3 hadoop hadoop 8914099 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-173-1287 -rw-r--r-- 3 hadoop hadoop 5773258 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-173-1288 -rw-r--r-- 3 hadoop hadoop 8950742 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-174-1287 -rw-r--r-- 3 hadoop hadoop 5829613 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-174-1288 -rw-r--r-- 3 hadoop hadoop 8808161 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-175-1287 -rw-r--r-- 3 hadoop hadoop 5910085 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-175-1288 -rw-r--r-- 3 hadoop hadoop 8871508 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-176-1287 -rw-r--r-- 3 hadoop hadoop 5896191 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-176-1288 -rw-r--r-- 3 hadoop hadoop 8855378 2020-11-26 08:59 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-177-1287 -rw-r--r-- 3 hadoop hadoop 5857271 2020-11-26 09:00 hdfs://xxx/test.db/hdfs_table/day=2020-11-26/hour=08/minute=59/part-633fa7d2-8d80-4e4d-ab90-f8b77c792b40-177-1288

*来自志愿者整理的flink邮件归档

展开
收起
说了是一只鲳鱼 2021-12-06 11:43:47 1329 0
1 条回答
写回答
取消 提交回答
  • Hi, admin 结合这个 issue 和你的对比结果, 我觉得应该是这个bug,这个问题在最新的分支已经修复,今天社区cut branch了,你可以帮忙在1.12的分支或master的分支上验证下吗?

    祝好, Leonard [1] https://github.com/apache/flink/tree/release-1.12 https://github.com/apache/flink/tree/release-1.12 [2] https://github.com/apache/flink/tree/master https://github.com/apache/flink/tree/master

    *来自志愿者整理的flink邮件归档

    2021-12-06 14:10:40
    赞同 展开评论 打赏
问答排行榜
最热
最新

相关电子书

更多
海量数据分布式存储——Apache HDFS之最新进展 立即下载
Hive Bucketing in Apache Spark 立即下载
2019大数据技术公开课第五季—Hive迁移到MaxCompute最佳实践 立即下载

相关实验场景

更多