You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hadoop.apache.org by Kellen Arb <ka...@comprehend.com> on 2017/09/07 17:46:20 UTC
When is an hdfs-* service restart required?
Hello,
I have a seemingly simple question, to which I can't find a clear answer.
Which services/node-types must be restarted for each of the configuration properties? For example, if I update the 'dfs.datanode.handler.count' property in the `hdfs-site.xml` configuration file, which services must be restarted? Can I get away with only restarting `datanodes`, or do I also need to restart `journalnodes` and/or `namenodes`, `zkfc` etc.
Just looking for clarification on this issue, though it seems that generally the advice is "restart everything”?
Thank you,
Kellen Arb
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@hadoop.apache.org
For additional commands, e-mail: user-help@hadoop.apache.org
Re: When is an hdfs-* service restart required?
Posted by Ravi Prakash <ra...@gmail.com>.
Hi Kellen!
The first part of the configuration is a good indication of which service
you need to restart. Unfortunately the only way to be completely sure is to
read the codez. e.g. most hdfs configuration is mapped to variables in
DFSConfigKeys
$ find . -name *.java | grep -v test | xargs grep
"dfs.datanode.handler.count"
./hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java:
public static final String DFS_DATANODE_HANDLER_COUNT_KEY =
"dfs.datanode.handler.count";
Looking at where this is used:
$ find . -name *.java | grep -v test | xargs grep
DFS_DATANODE_HANDLER_COUNT_KEY
./hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java:import
static org.apache.hadoop.hdfs.DFSConfigKeys.DFS_DATANODE_HANDLER_COUNT_KEY;
./hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/DataNode.java:
getConf().getInt(DFS_DATANODE_HANDLER_COUNT_KEY,
./hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/DFSConfigKeys.java:
public static final String DFS_DATANODE_HANDLER_COUNT_KEY =
"dfs.datanode.handler.count";
Cheers,
Ravi
On Thu, Sep 7, 2017 at 10:46 AM, Kellen Arb <ka...@comprehend.com> wrote:
> Hello,
>
> I have a seemingly simple question, to which I can't find a clear answer.
>
> Which services/node-types must be restarted for each of the configuration
> properties? For example, if I update the 'dfs.datanode.handler.count'
> property in the `hdfs-site.xml` configuration file, which services must be
> restarted? Can I get away with only restarting `datanodes`, or do I also
> need to restart `journalnodes` and/or `namenodes`, `zkfc` etc.
>
> Just looking for clarification on this issue, though it seems that
> generally the advice is "restart everything”?
>
> Thank you,
> Kellen Arb
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@hadoop.apache.org
> For additional commands, e-mail: user-help@hadoop.apache.org
>
>
Re: When is an hdfs-* service restart required?
Posted by Mingliang Liu <li...@apache.org>.
Restarting datanode(s) only is OK in this case.
Thanks,
> On Sep 7, 2017, at 10:46 AM, Kellen Arb <ka...@comprehend.com> wrote:
>
> Hello,
>
> I have a seemingly simple question, to which I can't find a clear answer.
>
> Which services/node-types must be restarted for each of the configuration properties? For example, if I update the 'dfs.datanode.handler.count' property in the `hdfs-site.xml` configuration file, which services must be restarted? Can I get away with only restarting `datanodes`, or do I also need to restart `journalnodes` and/or `namenodes`, `zkfc` etc.
>
> Just looking for clarification on this issue, though it seems that generally the advice is "restart everything”?
>
> Thank you,
> Kellen Arb
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@hadoop.apache.org
> For additional commands, e-mail: user-help@hadoop.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@hadoop.apache.org
For additional commands, e-mail: user-help@hadoop.apache.org