-
Type: Bug
-
Status: Open (View Workflow)
-
Priority: Major
-
Resolution: Unresolved
-
Affects Version/s: 2.3.0
-
Fix Version/s: None
-
Component/s: None
-
Labels:
-
Story Points:3
We started to observe unstable flow throughput test results (mostly with 5 nodes and 4 neighbors, see attachments) since 1/22/2020.
The dropped throughput number was due to random (~5% chance) high response delay from one of the onos nodes e.g.
/onos$ python3 /home/sdn/onos/tools/test/bin/flow-tester.py -f 700 -n 4 -s 10.192.19.68 -s 10.192.19.67 -s 10.192.19.64 -s 10.192.19.63 -s 10.192.19.62 -j {"server": "10.192.19.68", "elapsed": 29} {"server": "10.192.19.67", "elapsed": 30} {"server": "10.192.19.64", "elapsed": 31} {"server": "10.192.19.62", "elapsed": 40} {"server": "10.192.19.63", "elapsed": 3261}
The regression doesn't seem to be coming from onos code changes as it shows the same result by reverting the code back to beginning of Jan.
More investigation needs to be done to figure out whether the problem is caused by any environment changes in the testing infrastructure.