Created 01-22-2016 07:57 AM
Created 02-02-2016 10:23 PM
Here is what I did to get this done:
I downloaded Hive-1.2.1 binary from Apache hive repo (http://apache.arvixe.com/hive/hive-1.2.1/) into one of my edge node. Set up local HIVE environment and set HADOOP_USER_CLASSPATH_FIRST=true. This allowed me the switch Hive version on my specific bash_profile setup.
Thanks,
Pramit
Created 01-22-2016 08:24 AM
Created 01-22-2016 11:39 AM
Created 01-22-2016 10:08 PM
That link give you access to latest version of Hive.
Created 01-22-2016 09:56 PM
@Ali Bajwa and @Neeraj Sabharwal - Thanks for your quick response! However, please allow me to clarify my requirements - I already have Hadoop cluster with Hive version 0.13. I have recently read the HDP blogpost-(http://hortonworks.com/blog/announcing-apache-hive-0-14/) and now trying to test 0.14 features (ACID property in particular). Therefore I only need the latest version of Hive, not entire Hadoop distribution. Also, I don't want to impact existing version of Hive, I just want an additional Hive version only for my user id. So that, other users are not effected.
When, I looked for the Hive download page (http://www.apache.org/dyn/closer.cgi/hive/), I am not seeing any version called "0.14". Is 1.2.1 is equivalent to 0.14 ? Also, this is Apache download link, therefore can this binary be directly applied to Hortonwork existing distribution? Or do I need to download Hortonworks specific Hive version?
Thanks in advance!
Regards,
Pramit
Created 01-23-2016 02:10 AM
@Pramit Mitra at this point Ambari does not support multiple versions of Hive on the same cluster. Hive 0.13 and 0.14 are old, the latest stable release is 1.2.1. If you want to test ACID, you need to use 1.2.1 as there are important fixes for ACID features. That said, you need to be on HDP 2.3 to take advantage of that. If you download our latest Sandbox, it has Hive 1.2.1 and you can take it for a test drive there.
Created 02-02-2016 09:23 PM
@Pramit Mitra are you still having problems with this? Can you provide your own solution or accept best answer?
Created 02-02-2016 10:23 PM
Here is what I did to get this done:
I downloaded Hive-1.2.1 binary from Apache hive repo (http://apache.arvixe.com/hive/hive-1.2.1/) into one of my edge node. Set up local HIVE environment and set HADOOP_USER_CLASSPATH_FIRST=true. This allowed me the switch Hive version on my specific bash_profile setup.
Thanks,
Pramit