failed to transfer file to ~/.ansible/tmp/ansible-tmp-xxx/setup.py: [Errno 2] No such file or directory

[avivian]

ISSUE TYPE

• Bug Report

COMPONENT NAME

ansible-playbook setup

ANSIBLE VERSION

ansible 2.2.1.0
  config file = /etc/ansible/ansible.cfg
  configured module search path = Default w/o overrides

CONFIGURATION

OS / ENVIRONMENT

Linux ip-x-x-x-x 3.16.0-4-amd64 #1 SMP Debian 3.16.7-ckt25-1 (2016-03-06) x86_64 GNU/Linux

SUMMARY

When running ansible localhost using cloud-init, the following error is given:

PLAY [localhost] ***************************************************************

TASK [setup] *******************************************************************
fatal: [localhost]: FAILED! => {"failed": true, "msg": "failed to transfer file to ~/.ansible/tmp/ansible-tmp-1487178101.89-244463273612786/setup.py: [Errno 2] No such file or directory: '~/.ansible/tmp/ansible-tmp-1487178101.89-244463273612786/setup.py'"}
        to retry, use: --limit @/tmp/ansible/server/deploy_rancher_master.retry

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

This works using version 2.2.0.0.

STEPS TO REPRODUCE

We run our playbooks on AWS as part of a cloud-init script used in a cloudformation template. The playbooks are downloaded on to the instance and run using:

ansible-playbook /tmp/ansible/server/deploy_rancher_master.yml --extra-vars "extra=vars"

Ansible is installed using apt:

apt-get -y -t jessie-backports install ansible

Sample cloudformation template can be provided if required.

EXPECTED RESULTS

Setup task to complete.

ACTUAL RESULTS

Error in setup task:
TASK [setup] *******************************************************************
fatal: [localhost]: FAILED! => {"failed": true, "msg": "failed to transfer file to ~/.ansible/tmp/ansible-tmp-1487178101.89-244463273612786/setup.py: [Errno 2] No such file or directory: '~/.ansible/tmp/ansible-tmp-1487178101.89-244463273612786/setup.py'"}
        to retry, use: --limit @/tmp/ansible/server/deploy_rancher_master.retry

This works using version 2.2.0.0

[jctanner]

@avivian please run ansbile-playbook with -vvvv and paste the results here.

needs_info

[BenBewickJet]

I've recently started having the same issue with ansible scripts called from /etc/rc.local

`Using /etc/ansible/ansible.cfg as config file
Loading callback plugin default of type stdout, v2.0 from /usr/lib/python2.7/dist-packages/ansible/plugins/callback/init.pyc

PLAYBOOK: startup.yml **********************************************************
1 plays in /etc/ansible/startup.yml

PLAY [all] *********************************************************************

TASK [gather ec2 facts] ********************************************************
task path: /etc/ansible/startup.yml:10
Using module file /usr/lib/python2.7/dist-packages/ansible/modules/core/cloud/amazon/ec2_facts.py
ESTABLISH LOCAL CONNECTION FOR USER: root
EXEC /bin/sh -c '( umask 77 && mkdir -p "echo ~/.ansible/tmp/ansible-tmp-1487627348.46-252760944418207" && echo ansible-tmp-1487627348.46-252760944418207="echo ~/.ansible/tmp/ansible-tmp-1487627348.46-252760944418207" ) && sleep 0'
PUT /tmp/tmpvmMpk_ TO ~/.ansible/tmp/ansible-tmp-1487627348.46-252760944418207/ec2_facts.py
fatal: [web]: FAILED! => {
"failed": true,
"msg": "failed to transfer file to /.ansible/tmp/ansible-tmp-1487627348.46-252760944418207/ec2_facts.py: [Errno 2] No such file or directory: '/.ansible/tmp/ansible-tmp-1487627348.46-252760944418207/ec2_facts.py'"
}`

I have discovered that it is creating the directory path as a literal ~ underneath the /etc/ansible

root@web:/etc/ansible/~/.ansible/tmp# pwd /etc/ansible/~/.ansible/tmp root@web:/etc/ansible/~/.ansible/tmp# ls -la total 16 drwx------ 4 root root 4096 Feb 21 08:49 . drwx------ 3 root root 4096 Feb 21 08:43 .. drwx------ 2 root root 4096 Feb 21 08:43 ansible-tmp-1487627033.76-180215344939948 drwx------ 2 root root 4096 Feb 21 08:49 ansible-tmp-1487627348.46-252760944418207
but then trying to use path with ~ is getting shell expansion to root home where the path does not actually exist.

Script runs correctly when run from a login session.

[pdecat]

Same issue here with 2.2.1 (ok with 2.2.0).

Two workarounds:

1. set those in ansible.cfg:

remote_tmp = $HOME/.ansible/tmp
local_tmp = $HOME/.ansible/tmp

2. pass those environment variables to ansible-local command:

ANSIBLE_LOCAL_TEMP=$HOME/.ansible/tmp
ANSIBLE_REMOTE_TEMP=$HOME/.ansible/tmp

Related to #20332 and #21339

Edit: as suggested by others, those workarounds only work when the HOME environment variable is defined, which is not the case with cloud-init. In that case, setting HOME=/root helps.

I can confirm what @pdecat said; 2.2.1 seems to have introduced this bug. Currently, I'm just installing ansible with pip install ansible==2.2.0 to avoid this bug.

[decafdennis]

We're running a small playbook with local connection triggered by a Serf member join/leave event, and that stopped working with 2.2.1. It works when running from an interactive shell with 2.2.1, and it works when run from Serf with 2.2.0. My workaround is to downgrade, like @jersten.

Error:

fatal: [10.0.11.79]: FAILED! => {"changed": true, "failed": true, "msg": "Source /root/.ansible/tmp/ansible-tmp-1489689949.12-21813754930704/source not found"}

[ansibot]

@avivian This issue is waiting for your response. Please respond or the issue will be closed.

click here for bot help