Text Blame History Raw

Common Operations to initialize a node/service in CentOS infra

Once the server (bare-metal, Virtual Machine - internal or EC2 instance- ) is deployed, we just need to add it to ansible inventory (probably already done already during deploy step so complete with the following information)

Requirements:

  • Machine is reachable
  • You have initial credentials (either already injected ssh key and sudo right or just other equivalent credentials)
  • Access to required Ansible inventory

Adding node

You first need to add the node into DNS (either internally or externally) so please have a look at the dedicated DNS section, and that means kicking the role-bind.yml or role-unbound.yml playbooks based on the need, and after having pushed the change to git.

Once the node is available, we need once to initialize the node to confirm access and ssh host key/fingerprint and then sign it with our SSH CA.

Let's start by first ensuring that we can log onto a node (in our example a EC2 instance):

ssh centos@artwork-1.dev.centos.org uptime
The authenticity of host 'artwork-1.dev.centos.org (<no hostip for proxy command>)' can't be established.
ECDSA key fingerprint is SHA256:TFnZOT68OAkUQdTm1kCwoPxEN8d/4v/kqinsPcFD/04.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added 'artwork-1.dev.centos.org' (ECDSA) to the list of known hosts.
 09:41:48 up 7 min,  0 users,  load average: 0.00, 0.06, 0.04

Now we can proceed with next steps. Let's add the node in the correct Ansible inventory (like for example CentOS-CI or CentOS-prod or CentOS-staging, etc.

We first need to ensure that usual sysadmins will be granted first the correct rights and we just need to add the node in ansible inventory first. If you already know which roles you want to directly apply feel free to add into correct [group_name] in the inventory or just add it to [unclassified] first. When done, we can just play manually (only once) some playbooks and from there machine will be automatically reconfigured when role/inventory is updated (see the Ansible section about this)

fqdn="artwork-1.dev.centos.org"
ansible-playbook playbooks/adhoc-grant-access.yml -u centos -l ${fqdn} # or add -k to ask for password if needed to inject ssh keys

Now that sysadmins have their keys injected (including yours), you can initialize the node , but it will create a temporary file that you can then copy into inventory for some gathered facts, so you can use --extras-vars just for this specific call

out_dir="/home/arrfab/ansible/out"
test -d ${out_dir} || mkdir -p ${out_dir}
ansible-playbook playbooks/adhoc-init-node.yml -l ${fqdn} --extra-vars "out_dir=${out_dir}"
cat ${out_dir}/${fqdn} >> inventory/host_vars/${fqdn}

The adhoc-init-node.yml will do the following :

  • (optional) retrieve the public IP and allow incoming connections from that new ip for some infra services restricted by iptables
  • retrieve ssh host keys, sign these and push the signed
  • retrieve locally some facts that can be used later for basic host_vars template
  • play the baseline role (common for all nodes but with different settings, based on inventory
  • (optional): play other roles that are tied to ansible inventory group membership (if you added the host already in some specific groups)

If you configured correctl

Now that machine is in ansible inventory, you can always add new role, based on group memberships, change settings through group_vars or host_vars, etc, so Ansible BAU