Documentation Request - How to install

[spuder]

It would be nice if the readme had a basic overview of how to quickly get up and running with kafka-manager.

[ghost]

pretty easy - I had a tough time with sbt stuff at first because of my firewall. Once I got that sorted I ran into another issue as I had another process listening on the default port 9000... In summary this is what I did:

git clone https://github.com/yahoo/kafka-manager.git sbt clean dist unzip ../kafka-manager/target/universal/kafka-manager-1.0-SNAPSHOT.zip to wherever /opt/kafka-manager-1.0-SNAPSHOT/bin/kafka-manager -Dhttp.port=9002 ( i had to use 9002 because of the conflicting port) then open the browser to :9092 and you are good

[spuder]

Thanks, that worked for a quick proof of concept. How would you set this up in a web server that could be used in production?

(Start automatically, log to file, ect..)

[mchesler]

I'm currently working on adding the sbt-native-packager plugin to generate an installable RPM that will provide wrapping like what you're looking for

[spuder]

That would be great. Currently I we use ubuntu, so a .deb would be nice. 'pleaserun' could make this a lot easier.

https://github.com/jordansissel/pleaserun

[prune998]

maintaining .deb and .rpm is a waste of time. Build , unzip and you're done...

[mchesler]

@spuder no worries -- the sbt-native-packager plugin will generate deb packages as well.

[spuder]

My goal is to install this with chef or run in a docker container. Packages would make this easier.

[patelh]

Yahoo has its own package manager so we build components differently and so we pointed everyone to the Play Framework documentation for deployment. We'll look into the native packager and add more documentation.

[spuder]

This issue can be closed once 3026f52 is merged in. I'd like to continue the conversation about packaging in issue #18

[rvantassell]

@prune998 - For those of us who maintain large environments and infrastructure (often with puppet, or similar), the RPM and DEB packages are much simpler than ZIP files. (ie. generally due to other repository management tools also in-use)

[prune998]

I do maintain large envs (using Ansible, really great, try it :) )

What I see is : using the repository of the distrib:

• add the repository (and its key)
• update
• install the package (maybe installed in /usr, /etc...)
• configure the installed application

This will require a "playbook" or RedHat style, one for Debian style, one for BSD style... and if you install it directly from the package (not using a repo), you kind of "break" all the useful things of the package system.

or you can :

• wget the .zip file
• unzip in /opt/whatever/apps/
• configure the application

If you then add the overhead for the devs to build packages and maintain a repo for every distrib....

Well, I will not be the one maintaining all this, but I seriously it have to be taken in account. On my side, a good config script and startup script would be a far better invested time for the devs.

[rvantassell]

Except @prune998, there's rpmbuild... once the spec file is built, building the RPM is no more difficult than creating the zip file; I'm not asking them to maintain a full repository... that's kind of insane (though I bet folks at Yahoo know how to do that, already).

On "our" side, there's just puppet... or simply (bare minimum, though this is not good/great puppet) . Read: Please don't actually try to use this, as there are better/cleaner and more complete ways to do this sort of thing... even though it'll "probably work." (YMMV)

class kafka-manager {
  package { 'kafka-manager':
    ensure => latest,
  }

  file { '/etc/kafka-manager.conf':
    content => template('kafka/kafka-manager.conf.erb'),
    owner   => 'root',
    group   => 'kafka-manager',
    mode    => '0640',
  }

  user { 'kafka-manager':
    ensure => present,
    shell => '/bin/false',
  }

  group { 'kafka-manager':
    ensure => present,
  }

  service { 'kafka-manager':
    ensure => 'running',
    enable => true,
  }

  User['kafka-manager'] ->
  Group['kafka-manager'] ->
  Package['kafka-manager'] ~>
  File['/etc/kafka-manager.conf'] ~>
  Service['kafka-manager']
}

Edit: Better line breaks in example puppet manifest.

[udhayamgit]

Anyone can tell me How to run this service in windows 10 , I'm getting the below exception C:\kafka-manager-master\target\universal\kafka-manager-1.3.3.14\bin>kafka-manager -Dhttp.port=9002 The input line is too long. The syntax of the command is incorrect.

[cyrillesaxo]

Answer for @udhayamgit
you will get that exception when the path is too long on windows. Solution: you need to move the directory to a shorter directory. For example move it to c:/

[chuttipayan]

Even removing the version helped solve the problem of input path being too long

[simonalexander2005]

I still have the issue with input line too long. To resolve it, I edited the .bat file and changed the line where "APP_CLASSPATH" is set, to the following:

set "APP_CLASSPATH=%APP_LIB_DIR%\..\conf\;%APP_LIB_DIR%\*"