# Strange AWS Problems

**URL:** <https://forum.ansible.com/t/strange-aws-problems/16151>\
**Category:** Ansible Project\
**Tags:** aws\
**Created:** [September 8, 2014, 9:10pm UTC](https://forum.ansible.com/t/strange-aws-problems/16151 "2014-09-08T21:10:52Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![Deniz\_Acay](https://avatars.discourse-cdn.com/v4/letter/d/bc8723/32.png) [@Deniz\_Acay](https://forum.ansible.com/u/Deniz_Acay)\
**Post date:** [September 8, 2014, 9:10pm UTC](https://forum.ansible.com/t/strange-aws-problems/16151/1 "2014-09-08T21:10:52Z")

</div>

Hi,

I had playbook that already worked with Rackspace cloud and recently we decided to migrate to AWS.

After changing rax module operations with ec2 counterparts, everything looked good at that moment. I am using the official Debian 7.6 image.

Here’s the weird thing, I can create the infrastructure needed and can SSH into the machine after that.

But for some reason, Ansible fails after few actions spanning two roles and gives me this error:

`{"failed": true, "parsed": false}`

“-vvvv” flag didn’t provide any useful information too and logs at the server indicates that the client is disconnecting for no reason.

I disabled the roles that were failing but later discovered that regardless of the actions, it always fails after a few actions at the same action. What is more weird is that sometimes it fails after performing 2 more actions.

Sorry if this is a duplicate but I couldn’t find any solutions for this strange behaviour and logs are not helpful. Any advice?

Regards,  
Deniz

---

<div class="post-metadata">

**Author:** ![Michael\_DeHaan1](https://avatars.discourse-cdn.com/v4/letter/m/5e9695/32.png) [@Michael\_DeHaan1](https://forum.ansible.com/u/Michael_DeHaan1)\
**Post date:** [September 8, 2014, 9:25pm UTC](https://forum.ansible.com/t/strange-aws-problems/16151/2 "2014-09-08T21:25:56Z")

</div>

Hi Dennis,

Can you share what version of Ansible you are running, and the output before and after this line?

Usually when Ansible hits a parse error it will share part of the traceback or error.

Seeing both the output and the playbook line that failed would be helpful.

Thanks!

---

<div class="post-metadata">

**Author:** ![Deniz\_Acay](https://avatars.discourse-cdn.com/v4/letter/d/bc8723/32.png) [@Deniz\_Acay](https://forum.ansible.com/u/Deniz_Acay)\
**Post date:** [September 12, 2014, 2:28pm UTC](https://forum.ansible.com/t/strange-aws-problems/16151/3 "2014-09-12T14:28:03Z")

</div>

I am running Ansible version 1.6.6.

I am not sure why but now output shows additional errors with debug2/debug3 levels.

Here’s the output of the last task which seems to be failing:

`

failed: [54.68.129.168] =\> {“failed”: true, “md5sum”: "d291e1a0289d0aee992726913

d86ecab", “parsed”: false}  
invalid output was:  
debug3: mux\_client\_read\_packet: read header failed: Broken pipe  
debug2: Received exit status from master 0  
Shared connection to 54.68.129.168 closed.

`

Task that is failing is:

`

- name: set the content of the time zone file  
copy: dest=/etc/timezone content=“{{ nqdo\_server\_timezone }}”

`

When I tried to run this play again, it failed in a different task.

I was on 3G when running this play and the debug messages are new, I will try it again with WiFi later and report if here are any differences.

Thanks

---

<div class="post-metadata">

**Author:** ![Michael\_DeHaan1](https://avatars.discourse-cdn.com/v4/letter/m/5e9695/32.png) [@Michael\_DeHaan1](https://forum.ansible.com/u/Michael_DeHaan1)\
**Post date:** [September 12, 2014, 3:39pm UTC](https://forum.ansible.com/t/strange-aws-problems/16151/4 "2014-09-12T15:39:39Z")

</div>

Sounds like an unreliable connection, though it’s curious it managed to return a mostly parseable response.

Also it would be a good idea to take this time to upgrade to the latest released version, which is 1.7.1