hive错误记录

---

此文来记录hive使用时的错误，不定时更新

1. Failed with exception java.io.IOException:org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-1167307331-192.168.6.100-1477735992829:blk_1073741922_1112 file=/project/sogou/raw/sogou.50w.utf8

   用dfs上建外部表时，执行查询时报错。

解决方法:

1.1 在浏览器上http://master:50070/corrupt_files.jsp 查看webUI

---

NameNode ‘master:9000’

Started:

Fri May 05 05:02:21 PDT 2017

Version:2.6.4, 5082c73637530b0b7e115f9625ed7fac69f937e6

Compiled:2016-02-12T09:45Z by jenkins from (detached from 5082c73)
Cluster ID:CID-00d33199-9e38-442e-852f-51d35856be87
Block Pool ID:BP-1167307331-192.168.6.100-1477735992829

Browse the filesystem

NameNode Logs

Go back to DFS home

---

Reported Corrupt Files

blk_1073741919 /project/sogou/ext/sogou.50w.utf8.flt
blk_1073741922 /project/sogou/raw/sogou.50w.utf8

Total: At least 2 corrupt file(s)

---

发现有2个 Corrupt Files

---

1 . 2 使用fsck 命令 。HDFS支持fsck命令用以检查各种不一致。fsck用以报告各种文件问题，如block丢失或缺少block等

    [hadoop@master hive]$  hadoop fsck /

DEPRECATED: Use of this script to execute hdfs command is deprecated.
    Instead use the hdfs command for it.

17/05/05 05:48:42 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Connecting to namenode via http://master:50070
FSCK started by hadoop (auth:SIMPLE) from /192.168.6.100 for path / at Fri May 05 05:48:44 PDT 2017
.....................................
/project/sogou/ext/sogou.50w.utf8.flt: CORRUPT blockpool BP-1167307331-192.168.6.100-1477735992829 block blk_1073741919

/project/sogou/ext/sogou.50w.utf8.flt: MISSING 1 blocks of total size 64403097 B..
/project/sogou/raw/sogou.50w.utf8: CORRUPT blockpool BP-1167307331-192.168.6.100-1477735992829 block blk_1073741922

/project/sogou/raw/sogou.50w.utf8: MISSING 1 blocks of total size 57403097 B.........................................
/user/hive/warehouse/userinfo/o.txt:  Under replicated BP-1167307331-192.168.6.100-1477735992829:blk_1073742062_1252. Target Replicas is 3 but found 2 replica(s).
Status: CORRUPT
 Total size:    124244773 B
 Total dirs:    116
 Total files:   78
 Total symlinks:        0
 Total blocks (validated):  73 (avg. block size 1701983 B)
  ********************************
  CORRUPT FILES:    2
  MISSING BLOCKS:   2
  MISSING SIZE:     121806194 B
  CORRUPT BLOCKS:   2
  ********************************
 Minimally replicated blocks:   71 (97.26028 %)
 Over-replicated blocks:    0 (0.0 %)
 Under-replicated blocks:   1 (1.369863 %)
 Mis-replicated blocks:     0 (0.0 %)
 Default replication factor:    2
 Average block replication: 1.9452055
 Corrupt blocks:        2
 Missing replicas:      1 (0.6802721 %)
 Number of data-nodes:      2
 Number of racks:       1
FSCK ended at Fri May 05 05:48:45 PDT 2017 in 54 milliseconds


The filesystem under path '/' is CORRUPT

---

1 . 3

发现正好有两个MISSING，说了有2个block坏了。
删除坏块 hadoop fsck -delete Corrupt Files.

删除后DFS上文件消失。需要重新上传文件。在重新建外部表，问题解决。