hadoop Safe mode is ON 的解决办法

安哥网络

搭了一个hadoop集群环境，近期总是出现读写文件错误的情况，查看name node的日志显示 （Safe mode is ON）

• Safe mode is ON. The ratio of reported blocks 0.2000 has not reached the threshold 0.9990. Safe mode will be turned off automatically.
  

再往上翻

• org.apache.hadoop.hdfs.server.namenode.SafeModeException: Cannot delete /home/hadoop/tmp/mapred/system. Name node is in safe mode.
  

google 解决方案有三种
1）执行命令：bin/hadoop dfsadmin -safemode leave
dfsadmin -safemode value 参数value的说明如下：
enter - 进入安全模式
leave - 强制NameNode离开安全模式
get -  返回安全模式是否开启的信息
wait - 等待安全模式结束。
2）重新格式化文件系统 =。=
3）bin/hadoop fsck /

第一种方法，需要每次都执行一遍，很纠结~
第二种方法，format的意思你懂得。。。。
第三种方法，如果数据多，那执行起来会很慢，没办法，慢慢等吧。

• hadoop@mhadoop:~/hadoop-1.0.0$ bin/hadoop fsck /
• FSCK started by hadoop from /192.168.1.110 for path / at Wed Mar 07 17:57:43 CST 2012
• .
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0001/job.jar: CORRUPT block blk_-8710406465335110666
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0001/job.jar: MISSING 1 blocks of total size 142465 B..
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0002/job.jar: CORRUPT block blk_8343735010139595189
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0002/job.jar: MISSING 1 blocks of total size 142465 B..
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0004/job.jar: CORRUPT block blk_4796918520991551976
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0004/job.jar: MISSING 1 blocks of total size 142465 B..
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0005/job.jar: CORRUPT block blk_8732649167016211914
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0005/job.jar: MISSING 1 blocks of total size 142465 B..
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0006/job.jar: CORRUPT block blk_2190395249671123528
• /tmp/hadoop/tmp/mapred/staging/hadoop/.staging/job_201202290728_0006/job.jar: MISSING 1 blocks of total size 142465 B..
• /tmp/hadoop/tmp/mapred/system/jobtracker.info:  Under replicated blk_3010497672022840180_1511. Target Replicas is 2 but found 1 replica(s).
• .
• /user/hadoop/input/README.txt:  Under replicated blk_1278803782511847062_1510. Target Replicas is 2 but found 1 replica(s).
• ..
• /user/hadoop/output/_logs/history/job_201202290728_0007_1330473591943_hadoop_word+count: CORRUPT block blk_-165491155081655193
• /user/hadoop/output/_logs/history/job_201202290728_0007_1330473591943_hadoop_word+count: MISSING 1 blocks of total size 13581 B..
• /user/hadoop/output/_logs/history/job_201202290728_0007_conf.xml: CORRUPT block blk_359906738089887519
• /user/hadoop/output/_logs/history/job_201202290728_0007_conf.xml: MISSING 1 blocks of total size 20344 B..
• /user/hadoop/output/part-r-00000: CORRUPT block blk_-3883649133311907523
• /user/hadoop/output/part-r-00000: MISSING 1 blocks of total size 158081 B.Status: CORRUPT
• Total size:    905701 B
• Total dirs:    20
• Total files:   11
• Total blocks (validated):      10 (avg. block size 90570 B)
•   ********************************
•   CORRUPT FILES:        8
•   MISSING BLOCKS:       8
•   MISSING SIZE:         904331 B
•   CORRUPT BLOCKS:       8
•   ********************************
• Minimally replicated blocks:   2 (20.0 %)
• Over-replicated blocks:        0 (0.0 %)
• Under-replicated blocks:       2 (20.0 %)
• Mis-replicated blocks:         0 (0.0 %)
• Default replication factor:    2
• Average block replication:     0.2
• Corrupt blocks:                8
• Missing replicas:              2 (100.0 %)
• Number of data-nodes:          1
• Number of racks:               1
• FSCK ended at Wed Mar 07 17:57:43 CST 2012 in 11 milliseconds
• The filesystem under path '/' is CORRUPT
• 
  摘自：http://blog.chinaunix.net/uid-233938-id-3124458.html