Running Hive Without Hadoop?

[David M <mc...@outlook.com>]

All,

I'm trying to run a Hiveserver2 instance locally with the latest code out of the repo to test out a few JIRAs. There is some documentation here on how to run Hive without Hadoop, but it is based on ant build protocols: https://cwiki.apache.org/confluence/display/Hive/DeveloperGuide#DeveloperGuide-RunningHiveWithoutaHadoopCluster

I've tried running mvn package, but when I try to run bin/hive, I get the following error:

Missing Hive Execution Jar: /root/hive/lib/hive-exec-*.jar

What command do I need to use with maven in order to build the binaries in a way that will let me run the binaries in the way I need to do some testing?

Thanks!

David McGinnis

RE: Running Hive Without Hadoop?

[David M <mc...@outlook.com>]

Sorry for the delay in getting back to you.

Really I'm just looking for what is the easiest way for me to test a change I'm making to the Hive source code, outside of automated tests? If I want to just compile my changes, and test it through beeline against a Hiveserver2 instance with my code in it, what is the best way to do that?

Your proposal wouldn't do that, but it did give me an idea. Running the command I gave initially gives me the tar balls, so I think I can just take one of those and follow what that docker file does with it instead. My network is causing me fits at the moment, but I'll try this as soon as I can.

I'm sure I'm not the only contributor to Hive who has wanted to do this, so I figured I'd see if anyone else had ideas of what they do to test your changes to Hive in a manual manner.

- -David

-----Original Message-----
From: Alan Gates <al...@gmail.com> 
Sent: Saturday, April 6, 2019 3:06 PM
To: dev@hive.apache.org
Subject: Re: Running Hive Without Hadoop?

It depends on what you mean by run without Hadoop.  You can run HS2 without a Hadoop cluster.  Take a look at
https://nam05.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Falanfgates%2Fsqltest%2Fblob%2Fmaster%2Fdbs%2Fhive%2Fv3_1%2FDockerfile&amp;data=02%7C01%7C%7Cbffcddd4ba8c420ecf7608d6bacb581a%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636901779842386507&amp;sdata=zr9%2F1DQ5Q0W6sWVFFlPfGGmcuGfBbpmlaD9VTDlHAq4%3D&amp;reserved=0
for an example Dockerfile.  But Hadoop code is still needed in the path.
Hive depends on Hadoop code for a number of things and cannot build without the Hadoop jars.

Alan.

On Fri, Apr 5, 2019 at 12:12 PM David M <mc...@outlook.com> wrote:

> All,
>
> I'm trying to run a Hiveserver2 instance locally with the latest code 
> out of the repo to test out a few JIRAs. There is some documentation 
> here on how to run Hive without Hadoop, but it is based on ant build protocols:
> https://nam05.safelinks.protection.outlook.com/?url=https%3A%2F%2Fcwik
> i.apache.org%2Fconfluence%2Fdisplay%2FHive%2FDeveloperGuide%23Develope
> rGuide-RunningHiveWithoutaHadoopCluster&amp;data=02%7C01%7C%7Cbffcddd4
> ba8c420ecf7608d6bacb581a%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C
> 636901779842396591&amp;sdata=TxhbE90V3tTTmTp1%2F%2FkRZ%2By9aLxNNFWbc8D
> 5gUYIaMo%3D&amp;reserved=0
>
> I've tried running mvn package, but when I try to run bin/hive, I get 
> the following error:
>
> Missing Hive Execution Jar: /root/hive/lib/hive-exec-*.jar
>
> What command do I need to use with maven in order to build the 
> binaries in a way that will let me run the binaries in the way I need 
> to do some testing?
>
> Thanks!
>
> David McGinnis
>

Re: Running Hive Without Hadoop?

[Alan Gates <al...@gmail.com>]

It depends on what you mean by run without Hadoop.  You can run HS2 without
a Hadoop cluster.  Take a look at
https://github.com/alanfgates/sqltest/blob/master/dbs/hive/v3_1/Dockerfile
for an example Dockerfile.  But Hadoop code is still needed in the path.
Hive depends on Hadoop code for a number of things and cannot build without
the Hadoop jars.

Alan.

On Fri, Apr 5, 2019 at 12:12 PM David M <mc...@outlook.com> wrote:

> All,
>
> I'm trying to run a Hiveserver2 instance locally with the latest code out
> of the repo to test out a few JIRAs. There is some documentation here on
> how to run Hive without Hadoop, but it is based on ant build protocols:
> https://cwiki.apache.org/confluence/display/Hive/DeveloperGuide#DeveloperGuide-RunningHiveWithoutaHadoopCluster
>
> I've tried running mvn package, but when I try to run bin/hive, I get the
> following error:
>
> Missing Hive Execution Jar: /root/hive/lib/hive-exec-*.jar
>
> What command do I need to use with maven in order to build the binaries in
> a way that will let me run the binaries in the way I need to do some
> testing?
>
> Thanks!
>
> David McGinnis
>