For bug 1234153 the rpc_response_timeout for Heat in the undercloud was increased from 60s to 600s. This was confirmed to be working at the time. However, I installed an undercloud from the 2016-01-26 puddle yesterday and the value is not set:

  [cloud-user@undercloud-20160126 ~]$ sudo grep rpc_response_timeout /etc/heat/heat.conf /usr/share/heat/heat-dist.conf
  /etc/heat/heat.conf:#rpc_response_timeout = 60

Although the patch appears to still be in place:

  [cloud-user@undercloud-20160126 ~]$ grep rpc_response_timeout /usr/share/instack-undercloud/puppet-stack-config/puppet-stack-config.yaml.template 
  heat::rpc_response_timeout: 600

It looks like the patch to the puppet module has become corrupted, probably by a bad merge:

  [cloud-user@undercloud-20160126 ~]$ grep rpc_response_timeout /usr/share/openstack-puppet/modules/heat/manifests/init.pp
  # [*rpc_response_timeout*]
      'DEFAULT/rpc_response_timeout'         : value => $rpc_response_timeout;

This is missing the actual addition of $rpc_response_timeout to the heat class from the original patch: https://review.openstack.org/#/c/197964/3/manifests/init.pp


The same thing has been reported by customers using 7.2 (bug 1305557) and may be contributing to problems seen by QE in testing 7.3 (bug 1299613, bug 1302880, bug 1290949).


Installed packages:
openstack-puppet-modules-2015.1.8-47.el7ost.noarch
instack-undercloud-2.1.2-37.el7ost.noarch