Namenode in Safemode

This post shows what is Safemode in Namenode and what are the configurations for the safemode. You will also see the commands available to enter and leave the safemode explicitly.

When the Namenode is started it loads the file system state into memory initially from the fsimage and then apply the recent edits using the edit log file.

Then Namenode waits for the Datanodes to report their blocks. Note that information about the location of the blocks is kept in memory by Namenode, it is not stored in any file. So Namenode needs time to get block location from Datanodes.

During this process of loading the file system using fsimage and edit log and Datanode sending the list of their blocks Namenode stays in Safemode.

Safemode in Hadoop

Safemode for the NameNode is essentially a read-only mode for the HDFS cluster, where it does not allow any modifications to file system or blocks.

It is important for Namenode to stay in Safemode during this whole process because not getting enough time to get the list of blocks from Datanode would result in Namenode start replicating the blocks prematurely though enough replicas already exist in the cluster.

Normally the NameNode leaves Safemode automatically after the DataNodes have reported that most file system blocks are available.

How does Namenode know when enough block reports have come to exit Safemode. For that you need to configure properties in hdfs-site.xml (If you don’t configure explicitly default values are used).

dfs.namenode.safemode.threshold-pct – Specifies the percentage of blocks that should satisfy the minimal replication requirement defined by dfs.namenode.replication.min. Values less than or equal to 0 mean not to wait for any particular percentage of blocks before exiting safemode. Values greater than 1 will make safe mode permanent. Default for this property is 99.9% which means Namenode exits the Safemode when 99.9% blocks in the file system meet the minimum replication level.

dfs.namenode.safemode.extension – Determines extension of safe mode in milliseconds after the threshold level is reached. Default is 30000 miliseconds (30 Seconds).

If the default values are taken then Namenode exits the Safemode when 99.9% blocks in the file system meet the minimum replication level plus an extension of 30 seconds.

There is also property for specifying the minimum number of Datanodes that are considered alive.

dfs.namenode.safemode.min.datanodes – Specifies the number of datanodes that must be considered alive before the name node exits safemode. Values less than or equal to 0 mean not to take the number of live datanodes into account when deciding whether to remain in safe mode during startup. Values greater than the number of datanodes in the cluster will make safe mode permanent. Default value for this property is 0.

HDFS commands for Safemode

Entering Safemode explicitly

HDFS could be placed in Safemode explicitly using following HDFS command

Leaving Safemode explicitly

If you want to force the Namenode to leave the Safemode. This command can be used when you get SafeModeException and the “Name node is in safe mode” message.

Checking the Safemode status in Hadoop

You can check the NameNode front page to see whether Safemode is on or off.

From command line you can use the following command to check that-

hdfs dfsadmin -safemode get

That’s all for the topic Namenode in Safemode. If something is missing or you have something to share about the topic please write a comment.


You may also like

2 Comments

  1. Pingback: HDFS High Availability - KnpCode

  2. Pingback: How to Compress Intermediate Output of The Map Phase - KnpCode

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.