[jira] [Created] (HDFS-16697) Randomly setting “dfs.namenode.resource.checked.volumes.minimum” will always prevent safe mode from being turned off

["Jingxuan Fu (Jira)" <ji...@apache.org>]

Jingxuan Fu created HDFS-16697:
----------------------------------

             Summary: Randomly setting “dfs.namenode.resource.checked.volumes.minimum” will always prevent safe mode from being turned off
                 Key: HDFS-16697
                 URL: https://issues.apache.org/jira/browse/HDFS-16697
             Project: Hadoop HDFS
          Issue Type: Bug
    Affects Versions: 3.1.3
            Reporter: Jingxuan Fu
            Assignee: Jingxuan Fu
             Fix For: 3.1.3


 
{code:java}
<property>
  <name>dfs.namenode.resource.checked.volumes.minimum</name>
  <value>1</value>
  <description>
    The minimum number of redundant NameNode storage volumes required.
  </description>
</property>{code}
 

We found that when setting the value of “dfs.namenode.resource.checked.volumes.minimum” is greater than the total number of storage volumes in the NameNode, it is always impossible to turn off the safe mode, and when in safe mode, the file system only accepts read data requests, but not delete, modify and other change requests, which is greatly limited by the function.

The default value of the configuration item is 1, we set to 2 as an example for illustration, after starting hdfs logs and the client will throw the relevant reminders.

 
{code:java}
2022-07-27 17:37:31,772 WARN org.apache.hadoop.hdfs.server.namenode.FSNamesystem: NameNode low on available disk space. Already in safe mode.
2022-07-27 17:37:31,772 INFO org.apache.hadoop.hdfs.StateChange: STATE* Safe mode is ON.
Resources are low on NN. Please add or free up more resourcesthen turn off safe mode manually. NOTE:  If you turn off safe mode before adding resources, the NN will immediately return to safe mode. Use "hdfs dfsadmin -safemode leave" to turn safe mode off.
{code}
 

 
{code:java}
org.apache.hadoop.hdfs.server.namenode.SafeModeException: Cannot create directory /hdfsapi/test. Name node is in safe mode.
Resources are low on NN. Please add or free up more resourcesthen turn off safe mode manually. NOTE:  If you turn off safe mode before adding resources, the NN will immediately return to safe mode. Use "hdfs dfsadmin -safemode leave" to turn safe mode off. NamenodeHostName:192.168.1.167
        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.newSafemodeException(FSNamesystem.java:1468)
        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.checkNameNodeSafeMode(FSNamesystem.java:1455)
        at org.apache.hadoop.hdfs.server.namenode.FSNamesystem.mkdirs(FSNamesystem.java:3174)
        at org.apache.hadoop.hdfs.server.namenode.NameNodeRpcServer.mkdirs(NameNodeRpcServer.java:1145)
        at org.apache.hadoop.hdfs.protocolPB.ClientNamenodeProtocolServerSideTranslatorPB.mkdirs(ClientNamenodeProtocolServerSideTranslatorPB.java:714)
        at org.apache.hadoop.hdfs.protocol.proto.ClientNamenodeProtocolProtos$ClientNamenodeProtocol$2.callBlockingMethod(ClientNamenodeProtocolProtos.java)
        at org.apache.hadoop.ipc.ProtobufRpcEngine$Server$ProtoBufRpcInvoker.call(ProtobufRpcEngine.java:527)
        at org.apache.hadoop.ipc.RPC$Server.call(RPC.java:1036)
        at org.apache.hadoop.ipc.Server$RpcCall.run(Server.java:1000)
        at org.apache.hadoop.ipc.Server$RpcCall.run(Server.java:928)
        at java.base/java.security.AccessController.doPrivileged(Native Method)
        at java.base/javax.security.auth.Subject.doAs(Subject.java:423)
        at org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1729)
        at org.apache.hadoop.ipc.Server$Handler.run(Server.java:2916){code}
 

According to the prompt, it is believed that there is not enough resource space to meet the corresponding conditions to close safe mode, but after adding or releasing more resources and lowering the resource condition threshold "dfs.namenode.resource.du.reserved", it still fails to close safe mode and throws the same prompt .

According to the source code, we know that if the NameNode has redundant storage volumes less than the "dfs.namenode.resource.checked.volumes.minimum" set the minimum number of redundant storage volumes will enter safe mode. After debugging, *we found that the current NameNode storage volumes are abundant resource space, but because the total number of NameNode storage volumes is less than the set value, so the number of NameNode storage volumes with redundancy space must also be less than the set value, resulting in always entering safe mode.*

In summary, it is found that the configuration item lacks a condition check and an associated exception handling mechanism, which makes it impossible to find the root cause of the impact when a misconfiguration occurs.

The solution I propose is to use Precondition.checkArgument() to check the value of the configuration item and give a prompt when the value is greater than the number of  NameNode storage volumes to avoid the misconfiguration from affecting the subsequent operation of the program.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org