Creating an OMERO server

Overview

Teaching: 20 min
Exercises: 15 min

Questions

Objectives

• Experience the installation of an OMERO server via Ansible

OME have published (and maintain) a set of example playbooks which use the same roles OME use in production which will build an OMERO server.

Here we’re going to execute one of those examples - the “public user” example.

Don’t worry!

Most of the following commands include some BASH shell expansions to simplify connecting to a Vagrant VM from different machines. It’s really not as complicated as this makes it look.

1 Acquire a copy of the examples (you may have done this already):

$ git clone https://github.com/ome/ansible-examples-omero.git 
# or
$ git clone git@github.com:ome/ansible-examples-omero.git

Change directory to the ‘public-user’ example in the clone of ansible-examples-omero

$ cd ansible-examples-omero/public-user

Install the prerequisite roles, defined in requirements.yml

$ ansible-galaxy install -r ../requirements.yml -p roles

Add some port forwarding to the Vagrantfile to allow us to connect to OMERO afterwards - and also remove the “provision” step - since we’re learning here how to do this ourselves.

$ rm Vagrantfile && wget https://gist.github.com/kennethgillen/648105ba0f78440ca41e45963c471744/raw/Vagrantfile

Want to manually edit the Vagrantfile instead?

We want to end up with the following: …

 config.vm.provider "virtualbox" do |vb|
 config.vm.network "forwarded_port", guest: 80, host: 8080
   auto_correct: true
 config.vm.network "forwarded_port", guest: 443, host: 8443
   auto_correct: true
 config.vm.network "forwarded_port", guest: 4063, host: 4063
   auto_correct: true
 config.vm.network "forwarded_port", guest: 4064, host: 4064
   auto_correct: true
   vb.customize....

Tell Vagrant to kick off the creation of our VM (Takes 30s)

$ vagrant up

Create an inventory file for the vagrant-driven local VM:

$ echo "localhost ansible_port=$(vagrant ssh-config | grep Port | awk '{print $2}') ansible_user=vagrant " > inventory-file

It should look similar to:

$ cat inventory-file

localhost ansible_port=2222 ansible_user=vagrant

Asking Vagrant how to connect to the virtual machine

Vagrant takes care of running >1 machine at once by assigning the VMs different port numbers for SSH for example. We can ask vagrant the SSH port of the current machine, which we need to give to the ansible-playbook command, via the host file. See docs.ansible.com for more. These shell snippets acheive the function of asking vagrant for these details.

What port is the current Vagrant VM using for SSH?

$ echo $(vagrant ssh-config | grep Port | awk '{print $2}')

Where is the SSH key used to connect to the current Vagrant VM’s vagrant user?

$ echo $(vagrant ssh-config | grep IdentityFile | awk '{print $2}')

Verify Ansible can connect to the VM - the setup module again.

ansible -i inventory-file -m setup localhost --private-key  $(vagrant ssh-config | grep IdentityFile | awk '{print $2}') 

Type yes when presented with the below:

The authenticity of host '[localhost]:2222 ([127.0.0.1]:2222)' can't be established.
RSA key fingerprint is 88:f6:8a:85:09:e1:32:33:2f:47:b3:70:b8:ae:77:b7.
Are you sure you want to continue connecting (yes/no)?

Type yes…

localhost | SUCCESS => {
    "ansible_facts": {
        "ansible_all_ipv4_addresses": [
            "10.0.2.15"
        ],
... 

Seeing ssh errors?

If you’ve an ~/.ssh/known_hosts entry for localhost then this new VM is likely to conflict. You will have to delete the existing entry with the corresponding line number in the output from Ansible. This can be done manually, or with the following snippets:

# Looking in the output for known_hosts:line-number
$ echo $(ansible -i inventory-file -m setup localhost --private-key  $(vagrant ssh-config | grep IdentityFile | awk '{print $2}')  | grep -oP '(?<=hosts:)\d+')
# Use sed to delete that line. Replace "LINENUMBER" with the line number from the output above.
$ sed -i.bak -e "LINENUMBERd" ~/.ssh/known_hosts

Now re-try the above Ansible connect attempt.

Putting it all together - running the ‘public user’ example playbook. Estimated 15m runtime.

$ ansible-playbook -i inventory-file --private-key  $(vagrant ssh-config | grep IdentityFile | awk '{print $2}') --become  playbook.yml

15m wait…

Now we should be able to connect to our local Vagrant-driven VM.

Open up a web browser, and connect to http://localhost:8080

If this port was already taken when we issued vagrant up, Vagrant should have auto-corrected and picked another port. We can ask for the current port mappings as follows:

$ vagrant port

The forwarded ports for the machine are listed below. Please note that
these values may differ from values configured in the Vagrantfile if the
provider supports automatic port collision detection and resolution.

    22 (guest) => 2222 (host)
  4063 (guest) => 4063 (host)
  4064 (guest) => 4064 (host)
    80 (guest) => 8080 (host)
   443 (guest) => 8443 (host)

Key Points

• OME publish example OMERO server playbooks

• Vagrantfile defines your VM, including NAT port mappings

• We can ask Vagrant about ports and ssh-keys at the CLI