-
Type:
Bug
-
Status: Closed (View Workflow)
-
Priority:
Major
-
Resolution: Done
-
Affects Version/s: None
-
Fix Version/s: 1.6.0
-
Component/s: None
-
Labels:
-
Story Points:3
-
Epic Link:
-
Sprint:Goldeneye Black Sprint #3, Goldeneye Black Sprint #4
When some installed intents cannot be satisfied ( e.g. intents to/from one device which is either brought down or isolated from the rest of the topology by link-down), they bounce between withdrawing and failed state.
More importantly, sometimes this issue is causing other intents that can be rerouted stuck in installing state for a long time.
For instance, in a 10-node ring topology, when we install 10 intents which pass switch s1 (we call intent set A) and then bring down s1, the intents can be rerouted quickly. However, if we pre-install some other intents to/from s1 (intent set B), after s1 is down, states of intent set B will become unstable and intent set A cannot be rerouted for minutes or even longer.
In addition, when this happens, most of the time the intent states are inconsistent among onos nodes.
- relates to
-
ONOS-4589 if intent part of a batch fails, all intents associated in batch fail
-
- Closed
-