You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kylin.apache.org by skyyws <sk...@163.com> on 2017/06/16 12:17:28 UTC

Re: Some questions about Kylin2.0

For Q3, you can try to make soft links for both hdfs-site.xml and mapred-site.xml. 

2017-06-16 

skyyws 



发件人:"lxw" <lx...@qq.com>
发送时间:2017-06-13 11:41
主题:Some questions about Kylin2.0 
收件人:"dev"<de...@kylin.apache.org>
抄送:

Hi,All : 

   I have some questions about Kylin2.0, and my environment: 
    hadoop-2.6.0-cdh5.8.3 
    hbase-1.2.0-cdh5.8.3 
    apache-kylin-2.0.0-bin-cdh57 
    spark-2.1.0-bin-hadoop2.6 


Q1: Kylin2.0 not support Spark2.0? 
      
     find-spark-dependency.sh:  
     spark_dependency=`find -L $spark_home -name 'spark-assembly-[a-z0-9A-Z\.-]*.jar' .... 


Q2: I want to use Kylin2.0 without Spark Cubing, but failed. 


     kylin.sh: 
     function retrieveDependency() { 
         #retrive $hive_dependency and $hbase_dependency 
         source ${dir}/find-hive-dependency.sh 
         source ${dir}/find-hbase-dependency.sh 
         source ${dir}/find-hadoop-conf-dir.sh 
         source ${dir}/find-kafka-dependency.sh 
         source ${dir}/find-spark-dependency.sh 


     If not found spark dependencies, Kylin can not start : 
     
     [hadoop@hadoop10 bin]$ ./kylin.sh start 
     Retrieving hadoop conf dir... 
     KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current 
     Retrieving hive dependency... 
     Retrieving hbase dependency... 
     Retrieving hadoop conf dir... 
     Retrieving kafka dependency... 
     Retrieving Spark dependency... 
     spark assembly lib not found. 


     after modify kylin.sh “####source ${dir}/find-spark-dependency.sh”, Kylin start success .. 


Q3: Abount kylin_hadoop_conf_dir ? 
  
     I make some soft link under $KYLIN_HOME/hadoop-conf (core-site.xml、yarn-site.xml、hbase-site.xml、hive-site.xml),  
     and set "kylin.env.hadoop-conf-dir=/home/bigdata/kylin/current/hadoop-conf", when I execute ./check-env.sh,  


     [hadoop@hadoop10 bin]$ ./check-env.sh  
     Retrieving hadoop conf dir... 
    /home/bigdata/kylin/current/hadoop-conf is override as the kylin_hadoop_conf_dir 
    KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current 
    -mkdir: java.net.UnknownHostException: cdh5 
    Usage: hadoop fs [generic options] -mkdir [-p] <path> ... 
    Failed to create /kylin20. Please make sure the user has right to access /kylin20 


    My HDFS with HA, fs.defaultFS is "cdh5",when I don't set "kylin.env.hadoop-conf-dir", and use HADOOP_CONF_DIR, HIVE_CONF, HBASE_CONF_DIR from envionment variables (/etc/profile), it was correct.  
      
         
Best Regards! 
lxw 

Re: Some questions about Kylin2.0

Posted by ShaoFeng Shi <sh...@apache.org>.
For Q1, yes Kylin is running with Spark 1.6.3 now; Will upgrade to Spark
2.x in future releases.

For Q2, if you don't use Spark cubing, you can modify the start bash
scripts to disable the checking for Spark.

For Q3, can the host name "cdh5" be resolved on this machine?

2017-06-16 20:55 GMT+08:00 Alberto Ramón <a....@gmail.com>:

> About Q2:
> I'm agree with you, I think is a *issue*
> To start Kylin you must check exists one Source, one Engine and one Storage
> system
> (for example, is not necesary have Hive and Kafka)
>
> Example Spark
> <https://github.com/apache/kylin/blob/c38def7b53dae81f9fde0520b1fb27
> 0804dde728/build/bin/find-spark-dependency.sh#L38>
>
> On 16 June 2017 at 13:17, skyyws <sk...@163.com> wrote:
>
> > For Q3, you can try to make soft links for both hdfs-site.xml and
> > mapred-site.xml.
> >
> > 2017-06-16
> >
> > skyyws
> >
> >
> >
> > 发件人:"lxw" <lx...@qq.com>
> > 发送时间:2017-06-13 11:41
> > 主题:Some questions about Kylin2.0
> > 收件人:"dev"<de...@kylin.apache.org>
> > 抄送:
> >
> > Hi,All :
> >
> >    I have some questions about Kylin2.0, and my environment:
> >     hadoop-2.6.0-cdh5.8.3
> >     hbase-1.2.0-cdh5.8.3
> >     apache-kylin-2.0.0-bin-cdh57
> >     spark-2.1.0-bin-hadoop2.6
> >
> >
> > Q1: Kylin2.0 not support Spark2.0?
> >
> >      find-spark-dependency.sh:
> >      spark_dependency=`find -L $spark_home -name
> > 'spark-assembly-[a-z0-9A-Z\.-]*.jar' ....
> >
> >
> > Q2: I want to use Kylin2.0 without Spark Cubing, but failed.
> >
> >
> >      kylin.sh:
> >      function retrieveDependency() {
> >          #retrive $hive_dependency and $hbase_dependency
> >          source ${dir}/find-hive-dependency.sh
> >          source ${dir}/find-hbase-dependency.sh
> >          source ${dir}/find-hadoop-conf-dir.sh
> >          source ${dir}/find-kafka-dependency.sh
> >          source ${dir}/find-spark-dependency.sh
> >
> >
> >      If not found spark dependencies, Kylin can not start :
> >
> >      [hadoop@hadoop10 bin]$ ./kylin.sh start
> >      Retrieving hadoop conf dir...
> >      KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current
> >      Retrieving hive dependency...
> >      Retrieving hbase dependency...
> >      Retrieving hadoop conf dir...
> >      Retrieving kafka dependency...
> >      Retrieving Spark dependency...
> >      spark assembly lib not found.
> >
> >
> >      after modify kylin.sh “####source ${dir}/find-spark-dependency.sh”,
> > Kylin start success ..
> >
> >
> > Q3: Abount kylin_hadoop_conf_dir ?
> >
> >      I make some soft link under $KYLIN_HOME/hadoop-conf
> > (core-site.xml、yarn-site.xml、hbase-site.xml、hive-site.xml),
> >      and set "kylin.env.hadoop-conf-dir=/home/bigdata/kylin/current/
> hadoop-conf",
> > when I execute ./check-env.sh,
> >
> >
> >      [hadoop@hadoop10 bin]$ ./check-env.sh
> >      Retrieving hadoop conf dir...
> >     /home/bigdata/kylin/current/hadoop-conf is override as the
> > kylin_hadoop_conf_dir
> >     KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current
> >     -mkdir: java.net.UnknownHostException: cdh5
> >     Usage: hadoop fs [generic options] -mkdir [-p] <path> ...
> >     Failed to create /kylin20. Please make sure the user has right to
> > access /kylin20
> >
> >
> >     My HDFS with HA, fs.defaultFS is "cdh5",when I don't set
> > "kylin.env.hadoop-conf-dir", and use HADOOP_CONF_DIR, HIVE_CONF,
> > HBASE_CONF_DIR from envionment variables (/etc/profile), it was correct.
> >
> >
> > Best Regards!
> > lxw
> >
>



-- 
Best regards,

Shaofeng Shi 史少锋

Re: Some questions about Kylin2.0

Posted by Alberto Ramón <a....@gmail.com>.
About Q2:
I'm agree with you, I think is a *issue*
To start Kylin you must check exists one Source, one Engine and one Storage
system
(for example, is not necesary have Hive and Kafka)

Example Spark
<https://github.com/apache/kylin/blob/c38def7b53dae81f9fde0520b1fb270804dde728/build/bin/find-spark-dependency.sh#L38>

On 16 June 2017 at 13:17, skyyws <sk...@163.com> wrote:

> For Q3, you can try to make soft links for both hdfs-site.xml and
> mapred-site.xml.
>
> 2017-06-16
>
> skyyws
>
>
>
> 发件人:"lxw" <lx...@qq.com>
> 发送时间:2017-06-13 11:41
> 主题:Some questions about Kylin2.0
> 收件人:"dev"<de...@kylin.apache.org>
> 抄送:
>
> Hi,All :
>
>    I have some questions about Kylin2.0, and my environment:
>     hadoop-2.6.0-cdh5.8.3
>     hbase-1.2.0-cdh5.8.3
>     apache-kylin-2.0.0-bin-cdh57
>     spark-2.1.0-bin-hadoop2.6
>
>
> Q1: Kylin2.0 not support Spark2.0?
>
>      find-spark-dependency.sh:
>      spark_dependency=`find -L $spark_home -name
> 'spark-assembly-[a-z0-9A-Z\.-]*.jar' ....
>
>
> Q2: I want to use Kylin2.0 without Spark Cubing, but failed.
>
>
>      kylin.sh:
>      function retrieveDependency() {
>          #retrive $hive_dependency and $hbase_dependency
>          source ${dir}/find-hive-dependency.sh
>          source ${dir}/find-hbase-dependency.sh
>          source ${dir}/find-hadoop-conf-dir.sh
>          source ${dir}/find-kafka-dependency.sh
>          source ${dir}/find-spark-dependency.sh
>
>
>      If not found spark dependencies, Kylin can not start :
>
>      [hadoop@hadoop10 bin]$ ./kylin.sh start
>      Retrieving hadoop conf dir...
>      KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current
>      Retrieving hive dependency...
>      Retrieving hbase dependency...
>      Retrieving hadoop conf dir...
>      Retrieving kafka dependency...
>      Retrieving Spark dependency...
>      spark assembly lib not found.
>
>
>      after modify kylin.sh “####source ${dir}/find-spark-dependency.sh”,
> Kylin start success ..
>
>
> Q3: Abount kylin_hadoop_conf_dir ?
>
>      I make some soft link under $KYLIN_HOME/hadoop-conf
> (core-site.xml、yarn-site.xml、hbase-site.xml、hive-site.xml),
>      and set "kylin.env.hadoop-conf-dir=/home/bigdata/kylin/current/hadoop-conf",
> when I execute ./check-env.sh,
>
>
>      [hadoop@hadoop10 bin]$ ./check-env.sh
>      Retrieving hadoop conf dir...
>     /home/bigdata/kylin/current/hadoop-conf is override as the
> kylin_hadoop_conf_dir
>     KYLIN_HOME is set to /home/hadoop/bigdata/kylin/current
>     -mkdir: java.net.UnknownHostException: cdh5
>     Usage: hadoop fs [generic options] -mkdir [-p] <path> ...
>     Failed to create /kylin20. Please make sure the user has right to
> access /kylin20
>
>
>     My HDFS with HA, fs.defaultFS is "cdh5",when I don't set
> "kylin.env.hadoop-conf-dir", and use HADOOP_CONF_DIR, HIVE_CONF,
> HBASE_CONF_DIR from envionment variables (/etc/profile), it was correct.
>
>
> Best Regards!
> lxw
>