Hi folks,
For some reason a single host in an ELB pool was reported by the ec2_elb module as not belong to any ELBs, was dropped from the ELB, however since “changed" was False, the playbook never added it back and caused a minor interruption. The only rather suspicious clue I see is that there are two logs for the module running right before it reports no changes.
2014-03-27 18:50:24,486 p=24273 u=ds | TASK: [deploy | de-register instance from ELBs] ********************
2014-03-27 18:50:24,541 p=24273 u=ds | <127.0.0.1> REMOTE_MODULE ec2_elb instance_id=“i-3c58c60c" region="us-east-1" state="absent"
2014-03-27 18:50:41,508 p=24273 u=ds | <127.0.0.1> REMOTE_MODULE ec2_elb instance_id=“i-3c58c60c" region="us-east-1" state="absent"
2014-03-27 18:50:53,737 p=24273 u=ds | Result from run 1 is: {'attempts': 1, u'changed': False, u'ansible_facts': {u'ec2_elbs': }}
2014-03-27 18:50:53,742 p=24273 u=ds | ok: [server-01] => {"ansible_facts": {"ec2_elbs": }, "attempts": 1, "changed": false, "item": “"}
Anyone notice this before? This is ansible 1.4.x
Thx!
-cs