l3 prevent port deletion doesn't handle missing ports
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
neutron |
Fix Released
|
Low
|
Kevin Benton | ||
Juno |
Fix Released
|
Low
|
Kevin Benton |
Bug Description
The l3 code to check to see if a port can be deleted does not handle the case where the port id it is passed does not refer to a port that still exists. This throws an exception and makes the API behavior inconsistent when two concurrent requests come in to delete the same port.[1] This is inconsistent because if the port is concurrently deleted after the l3 check is done but before the delete attempt is made, no exception will be raised.[2]
1. This happens frequently when horizon deletes the subnet and immediately deletes the network afterwards. The dhcp agent will delete it's port on the subnet cleanup and may rip the port right before the delete_network call does it's auto port cleanup. The auto port cleanup will then hit a portnotfound exception which goes uncaught.
2. https:/
Changed in neutron: | |
importance: | Undecided → Low |
Changed in neutron: | |
milestone: | none → kilo-2 |
status: | Fix Committed → Fix Released |
Changed in neutron: | |
milestone: | kilo-2 → 2015.1.0 |
Fix proposed to branch: master /review. openstack. org/151788
Review: https:/