serverspec | RSpec tests for your servers | Continuous Deployment library

Your first question:

... all of the tutorials I see for Serverspec use should rather than expect. Is this because Serverspec hasn't been updated to use expectations?

No, it is mostly because the author of the Serverspec project has a preference for the "should" syntax, and that's why the Serverspec documentation continues to use it. He has explained here that:

I'm using should syntax instead of expect syntax because I think should syntax is more readable than expect syntax and I like it.

Using expect syntax is recommended way because adding should to every object causes failures when used with BasicObject-subclassed proxy objects.

But the one-liner syntax used with the examples in this page doesn't add should to any objects, so this syntax doesn't cause the above problems. That's why I'm using the one-liner should syntax.

Be aware that should and expect come from the rspec-expectations project, and the Serverspec author is correct "should" rather than "expect" is fine the way he is using it.

There is more info on the original motivation for the expect syntax from the Rspec author Myron Marston here.

Your second question:

... how would I write the test using expectations instead of should?

If you still want to use the expect syntax, just replace should everywhere with is_expected.to everywhere. This works fine:

I found a way to solve the problem, albeit a pretty hacky one. The key to this problem lay in how the container is finally released. When there are no longer any references pointing to the Docker instance it will be garbage collected and the container killed and deleted. However the object instance is held in a class level variable as a singleton in the base class. It would seem to me the only way to "reset" specinfra is to call the clear method inherited on the Docker class.

In the end the following solved the problem and the correct class is being used to run each spec.

So it appears you have conflicting packages installed from possibly 'odd' locations.

1:java-1.8.0-openjdk-devel-1.8.0.181-3.b13.el7_5.i686 is obsoleted by 2000:gd-jdk8-1.8.0_92-fcs.x86_64 is telling us that gd-jdk8 is providing the same thing as java-1.8.0-openjdk-devel but R doesn't like that name and wants a proper java and java-devel package.

First thing you need to do is disable all of these odd repos you have enabled and only use the epel base and updates repos.

If you navigate to; /etc/yum.repos.d/ you should find a folder full of files. open them up and change enabled=1 to enabled=0 which will disable them.

Now you should be able to yum erase gd-jdk8 and remove that along with any deps. Then you should be able to do yum install java-devel which should give you the correct version.

Once you have done that you will need to downgrade you R-core and libRmath packages as it appears from our chat you have some how managed to install newer versions of those that R also wants.

Do; yum downgrade R-core libRmath and say 'yes' to the deps. That should bring you down to the correct version you need.

Lastly do yum install R and with a bit of luck it should install for you!

It looks like a lot of packages my be provided by an internal repo or something, as you had a lot of things prefixed 'gd' which are not actually gd as people might expect! I'd have a word with your operations/IT team to see what's going on.

Lastly remember you disabled all those other repos - well you may wish to re-enable them now BUT... if you ever do a yum update make very sure you know what it's going to do as it might try and replace things with those 'gd' prefixed packages.

Yum did actually tell you what was wrong with the is obsoleted line. But granted digging into the deps requires a little more knowledge especially when your not sure what your looking at, and it's not a 'stock' install of CentOS!

I finally figured out what went wrong, and actually it's pretty simple:

Let's assume that we have a single Serverspec test case like this one:

Inspec is better than serverspec to do integration testing since it is more readable and helps to represent compliance as code. I use it with chef but Inspec is supported in puppet as well. You can refer the github link for examples

Thanks to Hubert above, I changed

this

I suspect that typo in the test is your enemy.

describe file('/opt/splunkforwarder/etc/apps/search/local/inputs.conf}') do

Try to remove that bracket and run test again.

It seems like you are missing the password for centos user.

Add this to spec_helper.rb:

A redirect severspec using curl command is like this: