Hdfs web_ui in-depth explanation, service startup log analysis, NN SNN relationship
Hdfs startup sequence:
1 、 namenode
2 、 datenode
3 、 secondnamenode
4 、 yarn
Webui:
Default is port 50070, dfsheath
Ui content explanation:
ClusterID: generated when namenode is formatted (hdfs namenode-format)
BlockpoolID: block pool
For federation and memory management use. )
Namenode Journal status:
It's the edit log.
Namenode storage:
Image file storage information
Launch the log file directory $HADOOP_HOME/logs:
X.log: log information for most applications through log4j logging
X.out records standard output and standard error logs, with a small number of records
Naming rules for log files:
[frame name-user name-process name-hostname-log format suffix]
Secondary namenode:
1. The client's operation on hdfs will be recorded in the edits file, that is, journal.
2. The content in namenode memory is fsp_w_picpath+edits.
3. After namenode's fsp_w_picpath and edite files are merged regularly, they are synchronized to secondary namenode.