Time out on Guard resource when launched from autoscale group in AWS on centos 7

[jslusher]

I'm trying to get an AMI with a recent chef converge to re-converge from an autoscale group with a user-data script. I've been able to converge an instance already running without incident. I'm getting an indefinite timeout when I run the converge from the user-data script. This is what's logged:

[2015-04-23T17:20:47+00:00] INFO: Processing firewalld_port[8300/tcp] action add (discogs-base::default line 65)
[2015-04-23T17:20:47+00:00] INFO: Processing execute[add port 8300/tcp to zone] action run (/var/chef/cache/cookbooks/firewalld/providers/port.rb line 8)
[2015-04-23T17:20:47+00:00] INFO: Processing execute[Guard resource] action run (dynamically defined)

I tried adding a "sleep 60" prior to the converge, but it doesn't help.

[jslusher]

Here are a few more lines from the debug output of chef-client:

[2015-04-23T18:34:11+00:00] DEBUG: providers for generic execute resource enabled on node include: [Chef::Provider::Execute]
[2015-04-23T18:34:11+00:00] DEBUG: providers that refused resource execute[Guard resource] were: []
[2015-04-23T18:34:11+00:00] DEBUG: providers that support resource execute[Guard resource] include: [Chef::Provider::Execute]
[2015-04-23T18:34:11+00:00] DEBUG: providers that survived replacement include: [Chef::Provider::Execute]