java - What is differene between secondary namenode and passive namenode -
what other namenode use in highavailabilty of namenode known , other namenodes used in hadoop federation confused
hope information solves confusion
1)namenode::
namenode holds meta data hdfs namespace information, block information etc. namenode stores hdfs filesystem information in file named fsimage. updates file system (add/remove blocks) not updating fsimage file , appended edit log, fsimage contains mapping of blocks files , other file system properties called snapshot of namenode. when restaring, namenode reads fsimage , applies changes log file bring filesystem state date in memory.
2)secondary namenode::
the secondary namenode periodically pulls these two(edits nd fsimage) files, , namenode starts writing changes new edits file. then, secondary namenode merges changes edits file old snapshot fsimage file , creates updated fsimage file. updated fsimage file copied namenode.
3)failover namenode or passive namenode(only if ha enabled):: prior hadoop 2.0.0, namenode single point of failure (spof) in hdfs cluster. each cluster had single namenode, , if machine or process became unavailable, cluster whole unavailable until namenode either restarted or brought on separate machine. above problems addressed providing option of running 2 redundant namenodes in ha cluster, standby/passive namenode performs checkpoints of namespace state, , not necessary run secondary namenode, checkpointnode, or backupnode in ha cluster. in fact, error.
Comments
Post a Comment