copy fails silently on remote host on small $HOME/.ansible/tmp

[Torsten Reinhard]

Hi,

we are using Ansible 1.8.2, and we stumbled upon a problem after copying and unarchiving a (large) file:

 failed: [myhost.com] => {"failed": true}
msg: Failed to find handler to unarchive. Make sure the required command to extract the file is installed.
msg: Failed to find handler to unarchive. Make sure the required command to extract the file is installed.

The task(s) looks like:

- name: copy artifact to the target machine
  copy: src={{ local_nexus_jar.dest }} dest={{ remote_transfer_dir }}

- name: prepare | unarchive deployment package on target node directly to target directory
  unarchive:
    copy='no'
    src="{{remote_transfer_dir}}/{{wls_artifactId}}-{{wls_version}}.{{wls_type}}"
    dest="/opt/wls/"

The file we´ve previously downloaded has these attributes:

ok: [myhost.com] => {
 ...
  "local_nexus_jar": { "package": "tar.gz", "proxy": null, "size": 849995522, 
 ... 

After transfering it to the remote host it looks like:

 changed: [myhost.com] => {"changed": true, ...}

 dest: /opt/wls/transfer/wls-patched-12.1.3.0.0-SNAPSHOT.tar.gz
...
size: 629563392
src: /home/wls_de/.ansible/tmp/ansible-tmp-1435185465.25-138644251456481/source
state: file
uid: 507

After wondering a while we 1st found out that the copied file is corrupted (see filesize above)

Than we found out why:
The 
$HOME/.ansible/tmp

had´nt have enough disk space due to limited $HOME directory in our environments !

I´m wondering why:
* the copy task doesn´t fail due to "out of disk space" ?
* why do we see "Failed to find handler to unarchive" instead of I/O Error due to corrupted archive here ?

We worked around that problem by using synchronize Module - but an explanation would be nice, additionally a hint in the documentation of the copy Module.

Thanx for any reply, 

Torsten

[Brian Coca]

copy should fail, that is a bug, please open a PR

just curious, why not let unarchive copy the file?


--
Brian Coca