journalnode Can't scan a pre-transactional edit log异常处理

一个测试环境hadoop集群由于磁盘满导致宕机，启动后发现journalnode报如下异常：

2018-03-19 20:48:04,817 WARN  namenode.FSImage (EditLogFileInputStream.java:scanEditLog(359)) - Caught exception after scanning through 0 ops from /data1_4T/journal/mycluster/current/edits_inprogress_0000000000024973700 while determining its valid length. Position was 1011712

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)
        at org.apache.hadoop.hdfs.qjournal.server.Journal.scanStorageForLatestEdits(Journal.java:192)
        at org.apache.hadoop.hdfs.qjournal.server.Journal.<init>(Journal.java:152)
        at org.apache.hadoop.hdfs.qjournal.server.JournalNode.getOrCreateJournal(JournalNode.java:90)
        at org.apache.hadoop.hdfs.qjournal.server.JournalNode.getOrCreateJournal(JournalNode.java:99)
        at org.apache.hadoop.hdfs.qjournal.server.JournalNodeRpcServer.getEditLogManifest(JournalNodeRpcServer.java:189)
        at org.apache.hadoop.hdfs.qjournal.protocolPB.QJournalProtocolServerSideTranslatorPB.getEditLogManifest(QJournalProtocolServerSideTranslatorPB.java:224)
        at org.apache.hadoop.hdfs.qjournal.protocol.QJournalProtocolProtos$QJournalProtocolService$2.callBlockingMethod(QJournalProtocolProtos.java:25431)
        at org.apache.hadoop.ipc.ProtobufRpcEngine$Server$ProtoBufRpcInvoker.call(ProtobufRpcEngine.java:640)
        at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:982)
        at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2351)
        at org.apache.hadoop.ipc.Server$Handler$1.run(Server.java:2347)
        at java.security.AccessController.doPrivileged(Native Method)
        at javax.security.auth.Subject

【解决过程】

    namenode是无法启动的，发现JournalNode维护的edits文件损坏，3台JournalNode只有1台上的JournalNode的运行日志是正常的，初步判断这个JournalNode可用，我们把这个JournalNode的数据文件copy到其他两个上去

1、停集群服务

2、将JournalNode异常的2个节点上的数据文件备份移到别的目录，然后删除JournalNode数据文件，copy正常JournalNode的数据文件到这2台节点上来

3、修改copy过来的这些文件的权限

4、启动服务，恢复正常