mlimaloureiro / ansible-rackhd-inventory Goto Github PK

It was requested that dynamic inventories script should allow to filter by more than one group, i.e, retrieve the nodes that belong to two groups. e.g., We could want the nodes that belong to the groups "ceph-node" or "ceph-mon" and the group "new".
This change would allow us to append a new field in the config.yaml file named "filter_group" with a string value. We could go even further and change the way we define the "filter_group" to a list of strings to allow filtering by more than one group if needed be.

Acording to the documentation "If you intend to replace an existing inventory ini file with a dynamic provider, it must return a JSON object which contains an ‘all’ group that includes every host in the inventory as a member and every group in the inventory as a child. It should also include an ‘ungrouped’ group which contains all hosts which are not members of any other group."

What changes does this entail for our script ? At the moment we only have "_meta" and the "<group_name>" as outputs. I estimate that adding every host to the "all" attribute should not require much effort. However, are the "ungrouped" really necessary ?

At the moment the inventory script only allows the "--list" argument.
According to the documentation " Dynamic inventory providers must accept the --list and --host arguments."

According to this, we should also allow the --host argument.
Is the hostname also its Id ?
Should we retrieve the node's information with this call (GET /api/current/nodes/)?

One of the features requested was that the RackHD url and the path to the config.yaml file could be changed by environment variables. The RackHD url defined in the environment variable would take precedence over the one defined in the config.yaml file.

As the tittle suggest, if there are no groups specified in the config.yaml file we should retrieve all groups in the RackHD. However, from my readings of the RackHD api documentation, we can get the list of tags with GET "/api/current/tags" and then for each tag GET the nodes that belong to that group. This works, howbeit, it is very inefficient. We should get all the nodes in RackHD with one call and the groups that that node belongs to should be included in the objects retrieved. Does something like "/api/current/tags/nodes" or "/api/current/nodes/tags" exist ?