FAQ
Hi All,

I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
4.0.3)

I need to install Hive on the cluster.

Is there a option in CM to automatically install Hive?

Thanks

Search Discussions

  • AT at Sep 5, 2012 at 4:46 pm
    Mike,

    If you are installing using CM Hive should already be there. Did you select
    all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
    4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks
  • Mike at Sep 5, 2012 at 5:07 pm
    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks



    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
    4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks
  • Mark Schnegelberger at Sep 6, 2012 at 2:26 am
    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive isn't
    installed through or managed in Cloudera Manager as other services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation. In addition,
    here's a video entitled "Hadoop Training: Introduction to Hive" you may
    find useful - http://www.cloudera.com/resource/introduction-to-apache-hive/

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 11, 2012 at 9:28 pm
    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
    4.0.3)

    Just typed hive at command line and it logged into hive and showed me the
    "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside the
    hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want to
    access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive isn't
    installed through or managed in Cloudera Manager as other services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation. In addition,
    here's a video entitled "Hadoop Training: Introduction to Hive" you may
    find useful -
    http://www.cloudera.com/resource/introduction-to-apache-hive/

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera



    On Wed, Sep 5, 2012 at 1:01 PM, Mike <mike...@gmail.com <javascript:>>wrote:
    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 2:00 pm
    Could you please respond?

    Thanks
    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
    4.0.3)

    Just typed hive at command line and it logged into hive and showed me the
    "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside the
    hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want to
    access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive isn't
    installed through or managed in Cloudera Manager as other services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation. In addition,
    here's a video entitled "Hadoop Training: Introduction to Hive" you may
    find useful -
    http://www.cloudera.com/resource/introduction-to-apache-hive/

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Philip Zeyliger at Sep 12, 2012 at 2:53 pm
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists (or on
    cdh-user). In general, right now there's HiveServer, but it can only serve
    one query at a time, which may not work for your use cases. A HiveServer2
    is in the works, I believe, but isn't out quite yet. Beeswax, which is an
    application that's part of Hue, is a web app that lets users submit Hive
    queries. You might be interested in checking that out.

    Cheers,

    -- Philip
    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using CDH
    4.0.3)

    Just typed hive at command line and it logged into hive and showed me the
    "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside the
    hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want to
    access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**display/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**
    resource/introduction-to-**apache-hive/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 3:54 pm
    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single user
    Apache Derby database?
    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve one
    query at a time, which may not work for your use cases. *

    Thanks


    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger wrote:

    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists (or
    on cdh-user). In general, right now there's HiveServer, but it can only
    serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike <mike...@gmail.com <javascript:>>wrote:
    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed me
    the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside
    the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want to
    access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**display/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**
    resource/introduction-to-**apache-hive/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Philip Zeyliger at Sep 12, 2012 at 3:33 pm

    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single user
    Apache Derby database?
    No, the current HiveServer implementation only supports doing one query.
    It's intended more as a bridge between a script and Hive than it is for
    multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve one
    query at a time, which may not work for your use cases. *

    Thanks


    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger wrote:

    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists (or
    on cdh-user). In general, right now there's HiveServer, but it can only
    serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed me
    the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside
    the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want to
    access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ**ay/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso**
    urce/introduction-to-**apache-**hive/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did you
    select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 5:32 pm
    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using sqoop). I
    am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).

    Will I be having the same problem like the hive single user issue please?

    Thoughts?

    Thanks

    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger wrote:



    On Wed, Sep 12, 2012 at 8:31 AM, Mike <mike...@gmail.com <javascript:>>wrote:
    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single user
    Apache Derby database?
    No, the current HiveServer implementation only supports doing one query.
    It's intended more as a bridge between a script and Hive than it is for
    multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve one
    query at a time, which may not work for your use cases. *

    Thanks


    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger wrote:

    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists (or
    on cdh-user). In general, right now there's HiveServer, but it can only
    serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed me
    the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside
    the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want
    to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ**ay/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso**
    urce/introduction-to-**apache-**hive/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during installation
    process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did
    you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Philip Zeyliger at Sep 12, 2012 at 5:33 pm

    On Wed, Sep 12, 2012 at 10:32 AM, Mike wrote:

    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using sqoop). I
    am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue please?

    Thoughts?

    Thanks

    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger wrote:


    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single user
    Apache Derby database?
    No, the current HiveServer implementation only supports doing one query.
    It's intended more as a bridge between a script and Hive than it is for
    multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve one
    query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists
    (or on cdh-user). In general, right now there's HiveServer, but it can
    only serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop ( using
    CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed me
    the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server outside
    the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want
    to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ****ay/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso****
    urce/introduction-to-**apache-**hi**ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did
    you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 6:02 pm
    Thanks the quick response, Philip.

    One last question.
    As I got these records into the hbase table, now I want to report it (write
    aggregate queries on the data) and display it.

    With hase-thrift php/python script, I can get the dataset from the hbase
    table for the specific range using row_key and
    do the aggregation on that.

    That why I was looking at Hive. Do I have any other choice to accomplish
    this please?

    Thanks


    On Wednesday, September 12, 2012 1:33:13 PM UTC-4, Philip Zeyliger wrote:



    On Wed, Sep 12, 2012 at 10:32 AM, Mike <mike...@gmail.com <javascript:>>wrote:
    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using sqoop). I
    am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue please?

    Thoughts?

    Thanks

    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger wrote:


    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single user
    Apache Derby database?
    No, the current HiveServer implementation only supports doing one query.
    It's intended more as a bridge between a script and Hive than it is for
    multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve one
    query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists
    (or on cdh-user). In general, right now there's HiveServer, but it can
    only serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed
    me the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server
    outside the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I want
    to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such, hive
    isn't installed through or managed in Cloudera Manager as other services
    are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ****ay/CDH4DOC/Hive+**Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso****
    urce/introduction-to-**apache-**hi**ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did
    you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Philip Zeyliger at Sep 12, 2012 at 6:04 pm
    I don't have any suggestions here at the moment. If Hive works for you,
    spawn off a hive process as a subprocess (one per query).

    -- Philip
    On Wed, Sep 12, 2012 at 10:56 AM, Mike wrote:

    Thanks the quick response, Philip.

    One last question.
    As I got these records into the hbase table, now I want to report it
    (write aggregate queries on the data) and display it.

    With hase-thrift php/python script, I can get the dataset from the hbase
    table for the specific range using row_key and
    do the aggregation on that.

    That why I was looking at Hive. Do I have any other choice to accomplish
    this please?

    Thanks


    On Wednesday, September 12, 2012 1:33:13 PM UTC-4, Philip Zeyliger wrote:


    On Wed, Sep 12, 2012 at 10:32 AM, Mike wrote:

    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using sqoop).
    I am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue please?

    Thoughts?

    Thanks


    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger
    wrote:
    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single
    user Apache Derby database?
    No, the current HiveServer implementation only supports doing one
    query. It's intended more as a bridge between a script and Hive than it is
    for multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve
    one query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists
    (or on cdh-user). In general, right now there's HiveServer, but it can
    only serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed
    me the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server
    outside the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I
    want to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps
    please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such,
    hive isn't installed through or managed in Cloudera Manager as other
    services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ******ay/CDH4DOC/Hive+**
    Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso******
    urce/introduction-to-**apache-**hi****ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there. Did
    you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 6:07 pm
    How do I spawn off a hive process as a subprocess (one per query)?

    Thanks
    On Wednesday, September 12, 2012 2:04:46 PM UTC-4, Philip Zeyliger wrote:

    I don't have any suggestions here at the moment. If Hive works for you,
    spawn off a hive process as a subprocess (one per query).

    -- Philip

    On Wed, Sep 12, 2012 at 10:56 AM, Mike <mike...@gmail.com <javascript:>>wrote:
    Thanks the quick response, Philip.

    One last question.
    As I got these records into the hbase table, now I want to report it
    (write aggregate queries on the data) and display it.

    With hase-thrift php/python script, I can get the dataset from the hbase
    table for the specific range using row_key and
    do the aggregation on that.

    That why I was looking at Hive. Do I have any other choice to accomplish
    this please?

    Thanks


    On Wednesday, September 12, 2012 1:33:13 PM UTC-4, Philip Zeyliger wrote:


    On Wed, Sep 12, 2012 at 10:32 AM, Mike wrote:

    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using sqoop).
    I am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue
    please?

    Thoughts?

    Thanks


    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger
    wrote:
    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single
    user Apache Derby database?
    No, the current HiveServer implementation only supports doing one
    query. It's intended more as a bridge between a script and Hive than it is
    for multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve
    one query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing lists
    (or on cdh-user). In general, right now there's HiveServer, but it can
    only serve one query at a time, which may not work for your use cases. A
    HiveServer2 is in the works, I believe, but isn't out quite yet. Beeswax,
    which is an application that's part of Hue, is a web app that lets users
    submit Hive queries. You might be interested in checking that out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    Just typed hive at command line and it logged into hive and showed
    me the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server
    outside the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I
    want to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps
    please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such,
    hive isn't installed through or managed in Cloudera Manager as other
    services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ******ay/CDH4DOC/Hive+**
    Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso******
    urce/introduction-to-**apache-**hi****ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there.
    Did you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop
    ( using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Philip Zeyliger at Sep 12, 2012 at 6:48 pm
    Mike,

    We're off pretty far from scm-users here; you might ask the hive user list.
    http://php.net/manual/en/function.system.php is the way for PHP to call
    "hive -e 'my query'". Mind you, that may not be a good way of doing it,
    because the Hive query can take a long time, etc.

    -- Philip
    On Wed, Sep 12, 2012 at 11:07 AM, Mike wrote:

    How do I spawn off a hive process as a subprocess (one per query)?

    Thanks

    On Wednesday, September 12, 2012 2:04:46 PM UTC-4, Philip Zeyliger wrote:

    I don't have any suggestions here at the moment. If Hive works for you,
    spawn off a hive process as a subprocess (one per query).

    -- Philip

    On Wed, Sep 12, 2012 at 10:56 AM, Mike wrote:

    Thanks the quick response, Philip.

    One last question.
    As I got these records into the hbase table, now I want to report it
    (write aggregate queries on the data) and display it.

    With hase-thrift php/python script, I can get the dataset from the hbase
    table for the specific range using row_key and
    do the aggregation on that.

    That why I was looking at Hive. Do I have any other choice to accomplish
    this please?

    Thanks


    On Wednesday, September 12, 2012 1:33:13 PM UTC-4, Philip Zeyliger wrote:


    On Wed, Sep 12, 2012 at 10:32 AM, Mike wrote:

    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using
    sqoop). I am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue
    please?

    Thoughts?

    Thanks


    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger
    wrote:
    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single
    user Apache Derby database?
    No, the current HiveServer implementation only supports doing one
    query. It's intended more as a bridge between a script and Hive than it is
    for multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve
    one query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing
    lists (or on cdh-user). In general, right now there's HiveServer, but it
    can only serve one query at a time, which may not work for your use cases.
    A HiveServer2 is in the works, I believe, but isn't out quite yet.
    Beeswax, which is an application that's part of Hue, is a web app that
    lets users submit Hive queries. You might be interested in checking that
    out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    Just typed hive at command line and it logged into hive and
    showed me the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server
    outside the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I
    want to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps
    please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such,
    hive isn't installed through or managed in Cloudera Manager as other
    services are.

    Please find the Installation and Configuration guide for Hive at
    https://ccp.cloudera.com/**displ********ay/CDH4DOC/Hive+**
    Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso******
    **urce/introduction-to-**apache-**hi******ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there.
    Did you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop
    ( using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera
  • Mike at Sep 12, 2012 at 6:47 pm
    Thanks for the quick responses, Philip.

    On Wednesday, September 12, 2012 2:41:55 PM UTC-4, Philip Zeyliger wrote:

    Mike,

    We're off pretty far from scm-users here; you might ask the hive user
    list. http://php.net/manual/en/function.system.php is the way for PHP to
    call "hive -e 'my query'". Mind you, that may not be a good way of doing
    it, because the Hive query can take a long time, etc.

    -- Philip

    On Wed, Sep 12, 2012 at 11:07 AM, Mike <mike...@gmail.com <javascript:>>wrote:
    How do I spawn off a hive process as a subprocess (one per query)?

    Thanks

    On Wednesday, September 12, 2012 2:04:46 PM UTC-4, Philip Zeyliger wrote:

    I don't have any suggestions here at the moment. If Hive works for you,
    spawn off a hive process as a subprocess (one per query).

    -- Philip

    On Wed, Sep 12, 2012 at 10:56 AM, Mike wrote:

    Thanks the quick response, Philip.

    One last question.
    As I got these records into the hbase table, now I want to report it
    (write aggregate queries on the data) and display it.

    With hase-thrift php/python script, I can get the dataset from the
    hbase table for the specific range using row_key and
    do the aggregation on that.

    That why I was looking at Hive. Do I have any other choice to
    accomplish this please?

    Thanks



    On Wednesday, September 12, 2012 1:33:13 PM UTC-4, Philip Zeyliger
    wrote:
    On Wed, Sep 12, 2012 at 10:32 AM, Mike wrote:

    Ok. One more question.

    Currently I have hbase table with 10 million rows (loaded using
    sqoop). I am also running thirft server on hbase.

    I am planning to install apache/php/thrift client on another server
    (outside hadoop) and write a php script to insert records into the above
    hbase table using thrift. This php script will be called from many
    clients(browser).
    The HBase Thrift server is fine for this, I believe.

    Will I be having the same problem like the hive single user issue
    please?

    Thoughts?

    Thanks


    On Wednesday, September 12, 2012 11:33:50 AM UTC-4, Philip Zeyliger
    wrote:
    On Wed, Sep 12, 2012 at 8:31 AM, Mike wrote:

    Hi Philip,

    Thanks for the response.

    Are you saying the below just because by default HIVE uses single
    user Apache Derby database?
    No, the current HiveServer implementation only supports doing one
    query. It's intended more as a bridge between a script and Hive than it is
    for multi-user stuff.

    -- Philip

    I can install mysql DB. Will that solve the problem?

    *> In general, right now there's HiveServer, but it can only serve
    one query at a time, which may not work for your use cases. *

    Thanks



    On Wednesday, September 12, 2012 10:53:25 AM UTC-4, Philip Zeyliger
    wrote:
    Hi Mike,

    You're best off asking about a Hive server on the Hive mailing
    lists (or on cdh-user). In general, right now there's HiveServer, but it
    can only serve one query at a time, which may not work for your use cases.
    A HiveServer2 is in the works, I believe, but isn't out quite yet.
    Beeswax, which is an application that's part of Hue, is a web app that
    lets users submit Hive queries. You might be interested in checking that
    out.

    Cheers,

    -- Philip

    On Wed, Sep 12, 2012 at 7:00 AM, Mike wrote:

    Could you please respond?

    Thanks

    On Tuesday, September 11, 2012 5:28:29 PM UTC-4, Mike wrote:

    Hi,

    I have 2 nodes (Centos) cluster running CM/hadoop/hbase/sqoop (
    using CDH 4.0.3)

    Just typed hive at command line and it logged into hive and
    showed me the "hive>" prompt.

    I did not do any installation. Looks like it came with CDH4.

    It is not managed thru CM but cluster installation includes hive.

    My need is:
    have apache/php/thrift client/mysql installed on the server
    outside the hadoop cluster.
    have thrift server service running on port 9090 in the hadoop cluster.
    have php script that accesses hbase thru thrift client. Now I
    want to access hive(hbase) in the php script also.

    Does hive have server / client like thrift? Any install steps
    please?

    Thanks

    On Wednesday, September 5, 2012 10:20:25 PM UTC-4, smark wrote:

    Hi Mike and AT,

    Hive is essentially a client-side app, not a service. As such,
    hive isn't installed through or managed in Cloudera Manager as other
    services are.

    Please find the Installation and Configuration guide for Hive
    at https://ccp.cloudera.com/**displ********ay/CDH4DOC/Hive+**
    Installation<https://ccp.cloudera.com/display/CDH4DOC/Hive+Installation>.
    In addition, here's a video entitled "Hadoop Training: Introduction to
    Hive" you may find useful - http://www.cloudera.com/**reso*****
    ***urce/introduction-to-**apache-**hi******ve/<http://www.cloudera.com/resource/introduction-to-apache-hive/>

    Best Regards,
    --
    Mark Schnegelberger
    Cloudera


    On Wed, Sep 5, 2012 at 1:01 PM, Mike wrote:

    I am using CM 4.0.3. I chose hadoop/hbase option during
    installation process because we wanted only hbase and hadoop.
    We even disabled hue / zooie process.

    Now I want to add Hive to the Cluster.

    Thanks




    On Wednesday, September 5, 2012 12:46:25 PM UTC-4, AT wrote:

    Mike,

    If you are installing using CM Hive should already be there.
    Did you select all the packages carefully during installation process?
    On Wednesday, September 5, 2012 7:09:17 AM UTC-7, Mike wrote:

    Hi All,

    I have 2 nodes (Centos) cluster running
    CM/hadoop/hbase/sqoop ( using CDH 4.0.3)

    I need to install Hive on the cluster.

    Is there a option in CM to automatically install Hive?

    Thanks

    --
    Mark Schnegelberger
    Customer Operations Engineer
    Cloudera

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedSep 5, '12 at 2:09p
activeSep 12, '12 at 6:48p
posts16
users4
websitecloudera.com
irc#hadoop

People

Translate

site design / logo © 2022 Grokbase