yarn - What's the right way to use historyserver of hadoop 2.2? -
I have been using hasoop-hasoop-2.2.0. Can I start history in master node and slave node?
-
But I'm not sure that I need a history server on the slave node?
-
If I start a history server on the master, can I get all the logs of all jobs?
-
If I have a command to start the master and slave node, do all the use of an order? Do not start each server one by one.
Any comments welcome.
You only need one historian to run it on any of the nodes you like, in which It also includes a dedicated node, but traditionally runs on the same node as a resource manager. A history server has been announced:
-
mapreduce.jobhistory.address
: MapReduce JobHistory server host: port is the default port 10020. -
mapreduce.jobhistory.webapp.address
: MapReduce JobHistory server web UI host: The port is the default port 19888. -
mapreduce.jobhistory.intermediate-done-dir
: The directory where the history files are written by Metrardous jobs (in HDFS) is the default/ mr-history / tmp < / Code>
-
mapreduce.jobhistory.done-dir
: Directory where history files are managed by MR JobHistory server HDFS). You can access through history, you do not directly use internal history files for casual browsing, history, default,/ mr-history / done
Resouremanager is available in the web UI.
Comments
Post a Comment