方式1
原因：namenode元数据被破坏，需要修复
解决：恢复一下namenode
hadoop namenode -recover
一路选择Y，一般就OK了
方式2
Need to copy the edits file to the journal node (I have 3 journal node s101 s102 s103)
[centos@s100 /home/centos/hadoop/ha/dfs/name1/current]$scp edits_0000000000000000041-0000000000000000043 centos@s101:/home/centos/hadoop/ha/dfs/journal/mycluster/current [centos@s100 /home/centos/hadoop/ha/dfs/name1/current]$scp edits_0000000000000000041-0000000000000000043 centos@s102:/home/centos/hadoop/ha/dfs/journal/mycluster/current [centos@s100 /home/centos/hadoop/ha/dfs/name1/current]$scp edits_0000000000000000041-0000000000000000043 centos@s103:/home/centos/hadoop/ha/dfs/journal/mycluster/current [centos@s100 /home/centos/hadoop/ha/dfs/name1/current]$hdfs zkfc -formatZK
https://blog.csdn.net/u011478909/article/details/51864071
https://blog.csdn.net/wqhaber/article/details/77529049
java.io.IOException: There appears to be a gap in the edit log. We expected txid ***, but got txid
原文：https://www.cnblogs.com/felixzh/p/9881036.html