Instead of waiting for up to a certain number of retries we set a high
timeout and only re-check every five seconds. Certain services can
take a minute or more to start and we want to avoid waisting resources
by polling too often.
@mpeters reported that we're not checking that the named service is
actually there after a reload. And that sometimes monit doesn't actually
return anything at all after a reload.
If there are already ongoing actions for a process managed by monit, the
module would exit unsuccessfully. It could also give off false positives
because it did not determine whether the service was started/stopped
when it was in a pending state. Which might be turning the service off,
but the action was to start it.
For example "Running - pending stop" would be regarded as the service
running and "state=enabled" would do nothing.
This will make Ansible wait for the state to finalize, or a timeout decided
by the new `max_retries` option, before it decides what to do.
This fixes issue #244.
* Remove leading module parameter on open_url call as it's no longer used
by module_utils.urls.open_url
* Force basic auth otherwise vsphere will just return a 401
If a bucket is being created in us-east-1, the module passed
'us-east-1' to boto's s3.create_bucket method rather than
Location.DEFAULT (an empty string). This caused boto to generate
invalid XML which AWS was unable to interpret.
This patch is adding a new module which allows to add and remove YUM
repository definitions. The module implements all repository options
as described in the `yum.conf` manual page.