Add undeploy commands #61
Conversation
tzumainn
force-pushed
the
undeploy
branch
2 times, most recently
from
4883df3 to
d276687
Compare
esiclient/v1/orchestrator/cluster.py
Outdated
| """Undeploy a cluster from ESI nodes""" | ||
|
|
||
| log = logging.getLogger(__name__ + ".Undeploy") | ||
| REQUIRED_FIELDS = ['nodes', 'private_network_name', 'api_vip', |
There was a problem hiding this comment.
I think REQUIRED_FIELDS is not used to check fields in config file in this class.
There was a problem hiding this comment.
Ah, good point - removed!
| if len(non_uuid_node_configs) > 0: | ||
| print("* %s node configs were skipped, as they do not" | ||
| " specify specific nodes" % len(non_uuid_node_configs)) | ||
| print(" * these nodes and ports will have to be removed" |
There was a problem hiding this comment.
How can user keep track of these unspecified nodes? Should we provide detailed instructions here?
There was a problem hiding this comment.
They'll just have to use existing CLI commands; there's no real good way to do this except openstack baremetal node list
| for node in node_uuids: | ||
| print(" * %s" % node) | ||
| ironic_client.node.set_provision_state(node, 'deleted') | ||
| print(" * waiting for nodes to start undeploy before" |
There was a problem hiding this comment.
Will you get an error if deleting a port while a node is still under deleting?
There was a problem hiding this comment.
Nope - but it takes ~10 seconds for Ironic to detach the port itself, so I figure there's no harm in waiting!
These undeploy commands allow users to cleanup orchestrated clusters.
|
@DanNiESh thanks for your comments! Answers in-line. Longer-term, I'm hoping to implement some changes to add some cluster metadata so that removing resources should be much easier. |
These undeploy commands allow users to cleanup orchestrated clusters.