Re: how to know the Spark worker Mechanism

[Yanbo Liang]

Did you set spark.executor.extraLibraryPath to the directory which your
native library exists?

2014-11-18 16:13 GMT+08:00 tangweihan tangwei...@360.cn:

 I'm a newbee in Spark. I know that what the work should do is written in
 RDD.
 But I want to make the worker load a native lib and I can do something to
 change the content of the lib in memory.
 So how can I do. I can do it on driver, but not worker. I always get a
 fatal
 error. The jvm report A fatal error
  C  [libstdc++.so.6+0x64d24]
 std::_Rb_tree_rotate_left(std::_Rb_tree_node_base*,
 std::_Rb_tree_node_base*)+0x4



 --
 View this message in context:
 http://apache-spark-user-list.1001560.n3.nabble.com/how-to-know-the-Spark-worker-Mechanism-tp19141.html
 Sent from the Apache Spark User List mailing list archive at Nabble.com.

 -
 To unsubscribe, e-mail: user-unsubscr...@spark.apache.org
 For additional commands, e-mail: user-h...@spark.apache.org

Re: how to know the Spark worker Mechanism

[tangweihan]

Ok. I don't put it in the path. Because this is not a lib I want to use
permanently. 
here is my code in RDD.
val fileaddr = SparkFiles.get(segment.so);
System.load(fileaddr);
val config = SparkFiles.get(qsegconf.ini)
val segment = new Segment//this is the native class
segment.init(config);//here failed if driver doesn't load this
lib

I just use the system.load to load this lib. But now I also  call some
functions in the lib to change some objects in the lib. It turns out the
fatal error. And after I first load it in driver, this works again in
standalone mode. I want to know how the job running from the driver to
worker, or how the worker memory is loaded to the native lib. Then I take
another sample lib. If the function not change the objects in the lib, the
lib can be only loaded in workers. And another thing I want to know is
whether there is something like the cache arche in hadoop streaming.



--
View this message in context: 
http://apache-spark-user-list.1001560.n3.nabble.com/how-to-know-the-Spark-worker-Mechanism-tp19141p19146.html
Sent from the Apache Spark User List mailing list archive at Nabble.com.

-
To unsubscribe, e-mail: user-unsubscr...@spark.apache.org
For additional commands, e-mail: user-h...@spark.apache.org