[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Issues all over with the comprehensive guide on an IAAS

----- Original Message -----
> From: "Kyle Crumpton (kcrumpto)" <kcrumpto cisco com>
> To: users lists openshift redhat com
> Sent: Thursday, September 26, 2013 3:15:17 PM
> Subject: Issues all over with the comprehensive guide on an IAAS
> Hi all,
> First off I'm a new user. 

Welcome! :)

> I'm following the comprehensive guide on an
> RHEL box using Rackspace for Infrastructure. (
> http://openshift.github.io/documentation/oo_deployment_guide_comprehensive.html
> )
> I keep getting to the point where I have a node host up and running
> with a broker and try to access the console to start deploying apps;
> however, the site displays that there was an error.
> I have found a few issues in the guide and am wondering if anyone
> else has come across these? Have been hitting a dead end trying to
> solve them.
> First issue is this command:
> yum install -y ruby193

The issues belong suggest to me that this didn't succeed. Can you check whether this RPM is installed? When I just did it, it resulted in 47 RPMs being installed, including ruby193-rubygem-bundler.

> cat <<EOF > /etc/profile.d/scl193.sh
> # Setup PATH, LD_LIBRARY_PATH and MANPATH for ruby-1.9
> ruby19_dir=$(dirname `scl enable ruby193 "which ruby"`)

What does this give you?
scl enable ruby193 "which ruby"

On my system it was:

> export PATH=$ruby19_dir:$PATH
> ruby19_ld_libs=$(scl enable ruby193 "printenv LD_LIBRARY_PATH")

Same question...

> export LD_LIBRARY_PATH=$ruby19_ld_libs:$LD_LIBRARY_PATH
> ruby19_manpath=$(scl enable ruby193 "printenv MANPATH")
> export MANPATH=$ruby19_manpath:$MANPATH

If the install of ruby193 (which is SCL-ized ruby version 1.9.3; system ruby is 1.8.7) didn't succeed, not much else is going to work. If the yum install above didn't work, there may be something wrong with your yum repositories. You should have RHEL, EPEL, openshift-dependencies, and openshift-origin. Review the repo setup sections... attach/pastebin your repos from /etc/yum.repos.d if you'd like us to take a look.

> cp -f /etc/profile.d/scl193.sh /etc/sysconfig/mcollective
> chmod 0644 /etc/profile.d/scl193.sh /etc/sysconfig/mcollective
> Setting the ruby19 dir and ld libs is basically a nil based on that
> install. Guide may be dated a bit? So I do end up getting Ruby193,
> but then when I install mcollective I run across another issue.
> Mcollective installs ruby version 1.8.7 and sets my ruby executable
> (/usr/bin/ruby) to that. I have tried to fix this using a link - "ln
> –sf /usr/bin/ruby193 /usr/bin/ruby" which upgrades my ruby version
> to 1.9.3.

Using the SCL ruby can be a little confusing, given you have two versions of ruby installed. It's about more than just ruby. Ruby has a whole ecosystem of tools - gem, bundler, rake, as well as libraries and gem locations, all of which must be kept separate for things to work out. If you want to work in a ruby 1.9.3 environment, you need to invoke the SCL, e.g.:

scl enable ruby193 ruby ...  (runs ruby 1.9.3 with all the libraries, gem locations, etc. correct)

Or do what I do:
scl enable ruby193 bash
(proceed to do whatever you needed to do in a ruby 1.9.3 context like you would normally)

It sounds like what you did below invoked system ruby 1.8 tools. That won't fly; the broker requires ruby 1.9.3.

> So I get to the point of bundling the broker plugins. Bundler does
> not come installed. So I do a: "gem install bundler" and then try to
> bundle the plugin directory.. I find out I need poltergeist v 1.2.0;
> I do a "gem install poltergeist –v 1.2.0" and come to find that I do
> not have gcc or gcc-c++ libs installed. I do a yum install of these
> libraries and execute again: "gem install bundler".. I find I need
> header directories (also not specified in the comprehensive guide).
> So I do a: "yum install –y ruby193-ruby-devel" at which point
> poltergeist will finally install and I can finally bundle the plugin
> directory. I go through and continue getting a node and console up
> and running. My only problem is by the end of this all I get a
> console up and running on HTTPD and tunnel through using my local
> machine. I find very vague errors on the page (no indication of what
> could cause them)
> Anyone have similar experiences?
> Thanks and BR
> Kyle
> _______________________________________________
> users mailing list
> users lists openshift redhat com
> http://lists.openshift.redhat.com/openshiftmm/listinfo/users

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]