-
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
-
Sprint:Goldeneye Sprint #1, Goldeneye Black Sprint #2
After running CHO (ATT topology) for two days, when we do the following loop:
install intents -> bring down link -> pingall -> bring up link -> uninstall intents
It is very frequent that intent rerouting after link down becomes slow. Flows are not updated for several minutes.
Specifically, e.g. after link s1-s2 down, intents are all in INSTALLED state and there are no pending_add flows, and paths is also updated, but flows toward s2 on s1 still point to the disabled port.
Some system information on ONOS nodes:
Note that in CHO test the system load keeps increasing along with the growing traffic (see attachment). Tshark summary shows the major traffic are from and to port 9876.
- is blocked by
-
ONOS-2322 ECMap: Support a tombstone expiration mechanism
-
- Closed
-