05-23-2016 10:37 AM - edited 05-23-2016 10:50 AM
I have a 2 Node YARN cluster using Cloudera manager 5.4.2 and by default sqoop 2 is installed. I wanted to use Sqoop 1 and I have installed Sqoop 1 client by following the cloudera documenatation as per below link:
Once the Sqoop 1 client is installed and I have deployed client configuration, restarted all services and also infact restarted entire cluster but still I see Sqoop 1 client is always in "None" state. Kindly Help
For reference, I have the screen capture from Cloudera Manager console for Sqoop 1:
05-24-2016 08:15 AM - edited 05-24-2016 10:17 AM
As the service name suggest 'Sqoop 1 Client', and not a service deamon with ability to report its state weather it's started or stopped. Are you seeing any issues with using sqoop command?
05-27-2016 03:55 AM
Thanks Michalis, am quite new to sqoop usage. I have installed Sqoop1 Client and Sqoop2 using Cloudera Manager. Am able to use "Sqoop2" command and it goes into sqoop prompt like below:
[hduser@node1 ~]$ sqoop2 Sqoop home directory: /opt/cloudera/parcels/CDH-5.4.1-1.cdh5.4.1.p0.6/lib/sqoop2 Sqoop Shell: Type 'help' or '\h' for help. sqoop:000>
But When I use, "Sqoop command", am getting the following message and doesn't enter into sqoop prompt:
[hduser@node1 ~]$ sqoop Warning: /opt/cloudera/parcels/CDH-5.4.1-1.cdh5.4.1.p0.6/bin/../lib/sqoop/../accumulo does not exist! Accumulo imports will fail. Please set $ACCUMULO_HOME to the root of your Accumulo installation. Try 'sqoop help' for usage. [hduser@node1 ~]$
Checked the sqoop version and it says 1.4.5 :
[hduser@node1 ~]$ sqoop version Warning: /opt/cloudera/parcels/CDH-5.4.1-1.cdh5.4.1.p0.6/bin/../lib/sqoop/../accumulo does not exist! Accumulo imports will fail. Please set $ACCUMULO_HOME to the root of your Accumulo installation. 16/05/27 16:20:13 INFO sqoop.Sqoop: Running Sqoop version: 1.4.5-cdh5.4.1 Sqoop 1.4.5-cdh5.4.1 git commit id Compiled by jenkins on Thu May 7 22:45:52 PDT 2015 [hduser@node1 ~]$
And I wanted to use Sqoop1 to import data from Oracle to HDFS/Hive. The reason I heard is Sqoop2 is not production ready and was advised to configure Sqoop1 as per architects.
Here am little confused, like which version of Sqoop am I using? Does Sqoop2 command also refers to sqoop1? How do I use Sqoop1 specifically?
05-30-2016 07:11 PM
06-01-2016 04:41 AM - edited 06-01-2016 04:42 AM
Thanks @Harsh J, Appreciate your help. Just as a feedback on part of Cloudera Documentation should also include an example with Sqoop1 and Sqoop2 and that would have been very helpful too for all the users who deploy Sqoop.