Intermittent connection refused errors using rax module

[Jimmy Prescott]

Hello,

Searched around before posting this but I was unable to find anything useful.

Anyway I am using the rax module to provision a server in the Rackspace Cloud using pyrax etc. The problem is that SOMETIMES the play will fail with a connection refused error.:

PLAY [Wait for port 22 to be ready] *******************************************

TASK: [wait_for port=22 delay=20] *********************************************
fatal: [testing6] => {'msg': 'FAILED: [Errno 111] Connection refused', 'failed': True}

FATAL: all hosts have already failed -- aborting

I tried adding the wait_for with a delay of 20 but maybe I am doing it wrong. Here is what my play looks like up to the point of failure:

---
- name: testing rackspace cloud
  hosts: localhost
  tasks:
    - name: Provision the cloud server
      local_action:
        module: rax
        name: "testing"
        flavor: "performance1-1"
        image: "centos-65"
        count: "1"
        group: "web"
        wait: yes
      register: rax

    - name: Add servers to 'raxhosts' group
      local_action:
        module: add_host
        hostname: "{{ item.name }}"
        ansible_ssh_host: "{{ item.rax_accessipv4 }}"
        ansible_ssh_pass: "{{ item.rax_adminpass }}"
        ansible_ssh_user: root
        groupname: raxhosts
      with_items: rax.success
      when: rax.action == 'create'

- name: Wait for port 22 to be ready
  hosts: raxhosts
  gather_facts: False
  tasks:
    - wait_for: port=22 delay=20

Any help on this is appreciated. The issue is intermittent sometimes it will run 10 times in a row then fail a few time then work again.

Thanks!

[Matt Martz]

I believe this email ended up being delayed a bit.  Jimmy and I worked through this yesterday and changed the wait_for task to look like:

- local_action: wait_for port=22 delay=20 hostname="{{ ansible_ssh_host }}"

-- 
Matt Martz
ma...@sivel.net

--
You received this message because you are subscribed to the Google Groups "Ansible Project" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ansible-proje...@googlegroups.com.
To post to this group, send email to ansible...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ansible-project/77da0480-8577-4699-961b-a37863ecfb79%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jimmy Prescott]

It hasn't failed since I made the change so thank you! Slight correction though in case anyone else runs into this problem. hostname="{{ ansible_ssh_host }}" is actually host="{{ ansible_ssh_host }}"

[Brian Spacecraft]

I see a similar issue probably 1/10 times I run a playbook I'm developing.

I am using wait_for after changing sshd listening port

- hosts: staging
  tasks:
    - name: Wait for ssh access on new port
      wait_for: delay=20 connect_timeout=30 host="{{ ansible_ssh_host }}" port="{{ ansible_ssh_port }}"

 And here is the error I get (I've substituted the actual port with <PORT>, and my username with <USER>):

GATHERING FACTS *************************************************************** 
<162.209.100.204> ESTABLISH CONNECTION FOR USER: <USER>
<162.209.100.204> REMOTE_MODULE setup
<162.209.100.204> EXEC ssh -C -tt -vvv -o ControlMaster=auto -o ControlPersist=60s -o ControlPath="/Users/brian/.ansible/cp/ansible-ssh-%h-%p-%r" -o StrictHostKeyChecking=no -o Port=<PORT> -o KbdInteractiveAuthentication=no -o PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey -o PasswordAuthentication=no -o User=<USER> -o ConnectTimeout=300 162.209.100.204 /bin/sh -c 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1432149978.13-279657059089785 && chmod a+rx $HOME/.ansible/tmp/ansible-tmp-1432149978.13-279657059089785 && echo $HOME/.ansible/tmp/ansible-tmp-1432149978.13-279657059089785'
fatal: [162.209.100.204] => SSH Error: ssh: connect to host 162.209.100.204 port <PORT>: Connection refused
    while connecting to 162.209.100.204:<PORT>

This probably doesn't have anything to do with the rax module. My playbook does the following:

1. Provision rackspace servers with the rax module
2. Apply a common role that as its last step reconfigures sshd to listen on a nonstandard port and only allow login by <USER>
3. Adds hosts to a new in-memory group (staging) with updated ansible_ssh_user and ansible_ssh_port
4. Runs the wait_for play above against these hosts

The error happens when it gathers facts for the wait_for play.

Any help would be much appreciated!