Instance fails to start when using a previously used port

[Itzik Brown]

An instance fails to start when it's attached with a previously port.

How to reproduce
================
1. Create a port
2. Launch an instance with the above port
3. Terminate the instance
4. Launch an instance with the above port
5. Verify that the instance is in 'spawning' state and then Error state

Version
========
python-networking-odl-11.0.0-0.20170721155128.a047729.el7ost.noarch
opendaylight-6.1.0-2.el7ost.noarch

[Isaku Yamahata (yamahata)]

IIRC when vm instance is terminated, attached ports are automatically deleted.
My memory can be wrong though.

[Isaku Yamahata (yamahata)]

What do you mean by terminate?
If it means really deleting VM, this bug is invalid.
If you mean something else, e.g. nova stop/start or suspend/resume, this report may make sense.
And also what's the port existence/status before re-starting the VM?

[Isaku Yamahata (yamahata)]

also Can you provide more concrete operation?
e.g. openstack server boot etc.

[Itzik Brown (itzikb1)]

By terminate I mean - delete. Why is it invalid ? With Neutron/OVS it works.
The status is unbound/DOWN

Steps:
# neutron net-create net1
# neutron subnet-create net1 192.168.99.1/24
# neutron port-create net1
# openstack server create --flavor m1.micro --image cirros --nic port-id=<port-id> vm1
# openstack server delete vm1
# openstack server create --flavor m1.micro --image cirros --nic port-id=<same-port-id> vm2

[Isaku Yamahata (yamahata)]

Can you please provide logs for it?

[Itzik Brown (itzikb1)]

Works with:
python-networking-odl-11.0.0-2.el7ost.noarch