insightdatascience/ansible-playbook

Ansible playbook to deploy distributed technologies

ansible-playbook's Issues

Error running the zookeeper.yml playbook

Issuing the command:

ansible-playbook ./zookeeper.yml --extra-vars "zookeeper_tag=Name_insight-test" --tags install

yields the following error:

PLAY [Creating host group zookeeper from dynamic inventory] *******************************************************************

TASK [Gathering Facts] ********************************************************************************************************
ok: [localhost]

TASK [add_host] ***************************************************************************************************************
fatal: [localhost]: FAILED! => {"failed": true, "msg": "'dict object' has no attribute u'tag_Name_insight_test'"}

NO MORE HOSTS LEFT ************************************************************************************************************
	to retry, use: --limit @/home/ubuntu/git/ansible-playbook/zookeeper.retry

PLAY RECAP ********************************************************************************************************************
localhost                  : ok=1    changed=0    unreachable=0    failed=1 

Here is the contents of my vars.yml file:

key_pair: ssh
instance_type: t2.medium
region: us-west-2
security_group_id: sg-7590d60e
num_instances: 2
subnet_id: subnet-1cd01347
tag_key_vals:
  Name: insight-test
  class: kafka
  class2: zookeeper

Any ideas if I'm doing something wrong or if a bug may be causing this?

Default inventory file location

Ubuntu 16.04, Ansible 2.3.0
As per the readme, a directory should be created at /etc/ansible/hosts. However, the default Ansible inventory location is /etc/ansible/hosts. This means the default inventory location specified in /etc/ansible/ansible.cfg must be changed to some other location. It would be good to specify it in the readme to not confuse Ansible newcomers. Thanks!

Tasks can't notify handlers defined after first one

Handlers after first one not recognized by ansible.

ERROR! handlers not found

Add `ssh-add <PEM_KEY>` instructions to README

It seems the pem key needs to be established with an auth agent for these scripts to work as designed.

FWIW, I think it'd be neat if we could find a configuration option that'd obviate the need for ssh-add, but it's fine as is.

Make Zookeeper a dependency for Kafka

Kafka stop command reports server is running even though it is stopped - info command verifies

Potentially a variable caching problem

EC2 `terminate` appears to succeed on failure to terminate ( reports `ok=2, changed=0` )

While nodes are initializing and running checks, the ansible ec2 terminate command succeeds (and the text is super-green!) even though the nodes are not terminating (as shown in the EC2 Management Console).

Later, after node initialization, the same terminate command shows nodes as changed in a yellow color, and nodes are actually shut down (according to AWS console).

changed=0 is fairly clear, but suggestion: report these as failed terminate commands.

Shorten the command length for playbook execution

1. Replace the --extra-vars command line args with a variable file for each playbook - kafka and zookeeper
2. Standardize variable file naming for playbooks so that playbooks look for a particular vars file and don't need command line args

A bug in experimental/Dockerfile

ansible-playbook/experimental/Dockerfile

Line 10 in 1b6a0f3

RUN apt-get update \

Hi, we are developing a tool to detect typos/bugs in dockerfiles automatically. Our tool finds a bug above. I think the problem is the last \ since it escapes the newline. This small typo occurs after you comment out the following apt-get install commands. I have tried it, and it outputs an error

The command '/bin/sh -c apt-get update RUN pip install boto' returned a non-zero code: 100

To fix it, I think you should either comment back the following apt-gets or delete the last \. Thank you for your confirmation. We are developing the tool, so your feedback will be greatly appreciated.

Multiple errors on EC2 launch

I think the primary need here is to update the README.

florp:ansible-playbook me$ ansible-playbook ./ec2.yml --extra-vars "vars_file=./example_ec2_vars.yml" --tags launch
[WARNING]: log file at /var/log/ansible.log is not writeable and we cannot create it, aborting

ERROR! Attempted to execute "/etc/ansible/hosts/ec2.py" as inventory script: Inventory script (/etc/ansible/hosts/ec2.py) had an execution error: b'Traceback (most recent call last):\n  File "/etc/ansible/hosts/ec2.py", line 128, in <module>\n    import boto\nImportError: No module named boto\n'

It seems ansible needs to be run as root in this configuration, is that right? I see I can change the /var/log/ansible.log entry in the ansible.cfg file, but does ansible need to run as root otherwise?

I installed ansible via pip3 (MacOS 10.12.3), but boto needed to be installed for python2.7. Is python2 explicitly required?