zoukankan      html  css  js  c++  java
  • hdp 集群问题解决记录

    2019-04-23 14:16:21,769 WARN namenode.FSImage (EditLogFileInputStream.java:scanEditLog(359)) - Caught exception after scanning through 0 ops from /hadoop/hdfs/journal/hnscluster/current/edits_inprogress_0000000000554042931 while determining its valid length. Position was 815104
    java.io.IOException: Can't scan a pre-transactional edit log.
    at org.apache.hadoop.hdfs.server.namenode.FSEditLogOp$LegacyReader.scanOp(FSEditLogOp.java:4974)
    at org.apache.hadoop.hdfs.server.namenode.EditLogFileInputStream.scanNextOp(EditLogFileInputStream.java:245)
    at org.apache.hadoop.hdfs.server.namenode.EditLogFileInputStream.scanEditLog(EditLogFileInputStream.java:355)
    at org.apache.hadoop.hdfs.server.namenode.FileJournalManager$EditLogFile.scanLog(FileJournalManager.java:551)

    原因:日志节点在日志中记录WARN以下,并且ambari警告日记网络ui无法访问
    解决:
    在有问题的节点上,将fsimage编辑目录(/hadoop/hdfs/journal/hnscluster/current)移动到备用位置。
    将fsimage edits目录(/ hadoop / hdfs / journal / stanleyhotel / current)从正常运行的JournalNode复制到此节点
    启动JournalNodes 或者启动hdfs


    under replicated blocks
    解决:
    找出没有复制的block:
    hdfs fsck / | grep 'Under replicated' | awk -F':' '{print $1}' >> /tmp/under_replicated_files
    然后循环修复:
    for hdfsfile in `cat /tmp/under_replicated_files`; do echo "Fixing $hdfsfile :" ; hadoop fs -setrep 3 $hdfsfile; done
    输出如下:
    Fixing /user/hdfs/.staging/job_1547173493660_0405/job.jar :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0405/job.jar
    Fixing /user/hdfs/.staging/job_1547173493660_0405/job.split :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0405/job.split
    Fixing /user/hdfs/.staging/job_1547173493660_0481/job.jar :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0481/job.jar
    Fixing /user/hdfs/.staging/job_1547173493660_0481/job.split :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0481/job.split
    Fixing /user/hdfs/.staging/job_1547173493660_0483/job.jar :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0483/job.jar
    Fixing /user/hdfs/.staging/job_1547173493660_0483/job.split :
    Replication 3 set: /user/hdfs/.staging/job_1547173493660_0483/job.split
    Fixing /user/hdfs/.staging/job_1547197402450_0021/job.jar :
    Replication 3 set: /user/hdfs/.staging/job_1547197402450_0021/job.jar
    Fixing /user/hdfs/.staging/job_1547197402450_0021/job.split :
    Replication 3 set: /user/hdfs/.staging/job_1547197402450_0021/job.split

  • 相关阅读:
    递归的效率问题以及递归与循环的比较
    malloc/free与new/delete的区别
    类的综合案例——纯虚函数与抽象类( 加强对接口与多态,以及派生类构造函数的理解 )
    对象移动、右值引用详解
    深入理解类成员函数的调用规则(理解成员函数的内存为什么不会反映在sizeof运算符上、类的静态绑定与动态绑定、虚函数表)
    为什么NULL指针也能访问成员函数?(但不能访问成员变量)
    洛谷P1656 炸铁路
    Vijos1901 学姐的钱包
    洛谷P2327 [SCOI2005] 扫雷
    洛谷P1993 小 K 的农场
  • 原文地址:https://www.cnblogs.com/shanhua-fu/p/10914269.html
Copyright © 2011-2022 走看看