# suggest having "allow\_destroy" tag in ec2\_module to prevent accidental deletion of servers

**URL:** <https://forum.ansible.com/t/suggest-having-allow-destroy-tag-in-ec2-module-to-prevent-accidental-deletion-of-servers/19261>\
**Category:** Ansible Project\
**Tags:** aws\
**Created:** [September 14, 2015, 4:05pm UTC](https://forum.ansible.com/t/suggest-having-allow-destroy-tag-in-ec2-module-to-prevent-accidental-deletion-of-servers/19261 "2015-09-14T16:05:27Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Kesten\_Broughton2](https://avatars.discourse-cdn.com/v4/letter/k/ecc23a/32.png) [@Kesten\_Broughton2](https://forum.ansible.com/u/Kesten_Broughton2)\
**Post date:** [September 14, 2015, 4:05pm UTC](https://forum.ansible.com/t/suggest-having-allow-destroy-tag-in-ec2-module-to-prevent-accidental-deletion-of-servers/19261/1 "2015-09-14T16:05:27Z")

</div>

I separate my creation and provisioning into aws.yml and site.yml.

Often we start with a few servers and then increase the size of a cluster later.  
I’d like to use aws exact\_count in the aws.yml so that it is idempotent but I do not want to allow destructive behavior.  
It would be nice if we could pass allow\_destroy to ec2\_module so that exact\_count would have “at least” semantics.

It can be pretty dangerous if you forget to limit the playbook correctly and accidentally delete a bunch of servers because your scope was too large.

I think non-destructive behaviour as a default would have avoided this situation as well.  
[https://github.com/ansible/ansible/issues/9431](https://github.com/ansible/ansible/issues/9431)

Would a pr for this be considered?

Kesten Broughton  
512 701 4209
