Hadoop Source code Interpretation Namenode High reliability: Ha;web way to view namenode information; dfs/data Decide Datanode storage location

Click Browserfilesystem. Same as command view results

When we look at the Hadoop source code, we see the Hdfs-default.xml file information under HDFs

We look for ${hadoop.tmp.dir} This is a reference variable, which is definitely defined in other files. As you can see in Core-default.xml, these two profiles have one thing in common:

Just do not change this file, but be able to copy information to Core-site.xml and hdfs-site.xml changes

Usr/local/hadoop is where I store my Hadoop directory.

A few key documents about Namenode

There is nothing in In_use.lock itself here. But it marks the namenode to be used. No other process calls

Important information is stored under current, especially fsimage. is Namenode core information

Edits for transaction processing

One of the HA measures is to back up multiple file information. Can be separated by commas to save multiple folders, note: Can not be changed directly in the Hdfs-default.xml. And should be copied to hdfs-site under change

The data file under DFS determines the Datanode storage location

Two HA measures: secondary for auxiliary namenode, merging edits to Fsimage.

Recovering data from secondary may be partially lost (like a Windows backup point). But I think the information here may be stored on the hard drive, and the information stored in the Namenode may also contain information in memory

The third of HA measures: replication. Set the number of replicas in the configuration file. There are different datanode on

Hadoop Source code Interpretation Namenode High reliability: Ha;web way to view namenode information; dfs/data Decide Datanode storage location