master-HA Full Network Partition

commit c3cdf9072df5869fb89b70af24eb3d3a028478fa (HEAD, origin/master, origin/HEAD, master)
Author: Thomas Vachuska [tom@onlab.us]
AuthorDate: Wed Apr 13 13:44:09 2016 -0700
Commit: Gerrit Code Review [gerrit@onlab.us]
CommitDate: Wed Apr 13 22:14:07 2016 +0000

Fixed the null-provider packet rate default value and ability to dynamically adjust the value.

Case 1: Setting up test environment - FAIL

Setup the test environment including installing ONOS, starting Mininet and ONOScli sessions.

• 1.1 Create cell file - No Result
• 1.2 Applying cell variable to environment - No Result
• 1.3 Starting Mininet - PASS
• 1.4 Git checkout and pull master - No Result
• 1.5 Using mvn clean install - PASS
• 1.6 Creating ONOS package - PASS
• 1.7 Installing ONOS package - FAIL
• ONOS install failed
• 1.8 Checking if ONOS is up yet - PASS
• 1.9 Checking ONOS nodes - PASS
• 1.10 Activate apps defined in the params file - No Result
• 1.11 Set ONOS configurations - PASS
• 1.12 App Ids check - PASS

Case 2: Assigning devices to controllers - PASS

Assign switches to ONOS using 'ovs-vsctl' and check that an ONOS node becomes the master of the device.

• 2.1 Assign switches to controllers - PASS

Case 8: Compare ONOS Topology view to Mininet topology - FAIL

Compare topology objects between Mininet and ONOS

• 8.1 Comparing ONOS topology to MN topology - FAIL
• ONOS topology don't match Mininet
• 8.2 Hosts view is consistent across all ONOS nodes - PASS
• 8.3 Hosts information is correct - PASS
• 8.4 Host attachment points to the network - PASS
• 8.5 Clusters view is consistent across all ONOS nodes - PASS
• 8.6 There is only one SCC - FAIL
• ONOS shows 6 SCCs
• 8.7 Device information is correct - FAIL
• Device information is incorrect
• 8.8 Links are correct - FAIL
• Links are incorrect
• 8.9 Hosts are correct - PASS
• 8.10 Checking ONOS nodes - PASS

Case 3: Adding host Intents - FAIL

Discover hosts by using pingall then assign predetermined host-to-host intents. After installation, check that the intent is distributed to all nodes and the state is INSTALLED

• 3.1 Install reactive forwarding app - PASS
• 3.2 Check app ids - PASS
• 3.3 Discovering Hosts( Via pingall for now ) - FAIL
• Reactive Pingall failed, one or more ping pairs failed
• 3.4 Uninstall reactive forwarding app - PASS
• 3.5 Check app ids - PASS
• 3.6 Add host intents via cli - FAIL
• Error looking up host ids
• 3.7 Intent Anti-Entropy dispersion - PASS

Case 4: Verify connectivity by sending traffic across Intents - FAIL

Ping across added host intents to check functionality and check the state of the intent

• 4.1 Check Intent state - PASS
• 4.2 Ping across added host intents - FAIL
• Intents have not been installed correctly, pings failed.
• 4.3 Check leadership of topics - PASS
• 4.4 Wait a minute then ping again - FAIL
• Intents have not been installed correctly, pings failed.

Case 5: Setting up and gathering data for current state - FAIL

• 5.1 Check that each switch has a master - FAIL
• Some devices don't have a master assigned
• 5.2 Get the Mastership of each switch from each controller - PASS
• 5.3 Check for consistency in roles from each controller - FAIL
• ONOS nodes have different views of switch roles
• 5.4 Get the intents from each controller - PASS
• 5.5 Check for consistency in Intents from each controller - PASS
• 5.6 Get the flows from each controller - PASS
• 5.7 Check for consistency in Flows from each controller - PASS
• 5.8 Get the OF Table entries - No Result
• 5.9 Start continuous pings - No Result
• 5.10 Collecting topology information from ONOS - No Result
• 5.11 Host view is consistent across ONOS nodes - PASS
• 5.12 Each host has an IP address - PASS
• 5.13 Cluster view is consistent across ONOS nodes - PASS
• 5.14 Cluster view correct across ONOS nodes - FAIL
• ONOS shows 6 SCCs
• 5.15 Comparing ONOS topology to MN - PASS
• 5.16 Device information is correct - FAIL
• Device information is incorrect
• 5.17 Links are correct - FAIL
• Links are incorrect
• 5.18 Hosts are correct - PASS

Case 14: Start Leadership Election app - FAIL

• 14.1 Install leadership election app - PASS
• 14.2 Run for election on each node - FAIL
• Inconsistent leaderboards

Case 16: Install Primitives app - PASS

• 16.1 Install Primitives app - PASS

Case 17: Check for basic functionality with distributed primitives - PASS

Test the methods of the distributed primitives (counters and sets) throught the cli

• 17.1 Increment then get a default counter on each node - PASS
• 17.2 Get then Increment a default counter on each node - PASS
• 17.3 Counters we added have the correct values - PASS
• 17.4 Add -8 to then get a default counter on each node - PASS
• 17.5 Add 5 to then get a default counter on each node - PASS
• 17.6 Get then add 5 to a default counter on each node - PASS
• 17.7 Counters we added have the correct values - PASS
• 17.8 Distributed Set get - PASS
• 17.9 Distributed Set size - PASS
• 17.10 Distributed Set add() - PASS
• 17.11 Distributed Set addAll() - PASS
• 17.12 Distributed Set contains() - PASS
• 17.13 Distributed Set containsAll() - PASS
• 17.14 Distributed Set remove() - PASS
• 17.15 Distributed Set removeAll() - PASS
• 17.16 Distributed Set addAll() - PASS
• 17.17 Distributed Set clear() - PASS
• 17.18 Distributed Set addAll() - PASS
• 17.19 Distributed Set retain() - PASS
• 17.20 Partitioned Transactional maps put - PASS
• 17.21 Partitioned Transactional maps get - PASS

Case 61: Partition ONOS nodes into two distinct partitions - PASS

• 61.1 Checking ONOS Logs for errors - No Result
• 61.2 Partitioning ONOS nodes - PASS

Case 8: Compare ONOS Topology view to Mininet topology - FAIL

Compare topology objects between Mininet and ONOS

• 8.1 Comparing ONOS topology to MN topology - FAIL
• ONOS topology don't match Mininet
• 8.2 Hosts view is consistent across all ONOS nodes - PASS
• 8.3 Hosts information is correct - PASS
• 8.4 Host attachment points to the network - PASS
• 8.5 Clusters view is consistent across all ONOS nodes - PASS
• 8.6 There is only one SCC - FAIL
• ONOS shows 6 SCCs
• 8.7 Device information is correct - FAIL
• Device information is incorrect
• 8.8 Links are correct - FAIL
• Links are incorrect
• 8.9 Hosts are correct - PASS
• 8.10 Checking ONOS nodes - PASS

Case 7: Running ONOS Constant State Tests - FAIL

• 7.1 Check that each switch has a master - FAIL
• Some devices don't have a master assigned
• 7.2 Read device roles from ONOS - PASS
• 7.3 Check for consistency in roles from each controller - FAIL
• ONOS nodes have different views of switch roles
• 7.4 Get the intents and compare across all nodes - PASS
• 7.5 Check for consistency in Intents from each controller - PASS
• 7.6 Compare current intents with intents before the failure - PASS
• 7.7 Get the OF Table entries and compare to before component failure - PASS
• 7.8 Leadership Election is still functional - FAIL
• Something went wrong with Leadership election

Case 4: Verify connectivity by sending traffic across Intents - FAIL

Ping across added host intents to check functionality and check the state of the intent

• 4.1 Check Intent state - PASS
• 4.2 Ping across added host intents - FAIL
• Intents have not been installed correctly, pings failed.
• 4.3 Check leadership of topics - PASS
• 4.4 Wait a minute then ping again - FAIL
• Intents have not been installed correctly, pings failed.

Case 15: Check that Leadership Election is still functional - FAIL

• 15.1 Run for election on each node - PASS
• 15.2 Check that each node shows the same leader and candidates - FAIL
• Nodes have different leaderboards
• 15.3 Find current leader and withdraw - FAIL
• Node was not withdrawn from election
• 15.4 Check that a new node was elected leader - FAIL
• Something went wrong with Leadership election
• 15.5 Check that that new leader was the candidate of old leader - FAIL
• Incorrect Candidate Elected
• 15.6 Run for election on old leader( just so everyone is in the hat ) - FAIL
• App failed to run for election
• 15.7 Check that oldLeader is a candidate, and leader if only 1 node - FAIL
• Something went wrong with Leadership election after the old leader re-ran for election

Case 17: Check for basic functionality with distributed primitives - FAIL

Test the methods of the distributed primitives (counters and sets) throught the cli

• 17.1 Increment then get a default counter on each node - PASS
• 17.2 Get then Increment a default counter on each node - PASS
• 17.3 Counters we added have the correct values - PASS
• 17.4 Add -8 to then get a default counter on each node - PASS
• 17.5 Add 5 to then get a default counter on each node - PASS
• 17.6 Get then add 5 to a default counter on each node - PASS
• 17.7 Counters we added have the correct values - PASS
• 17.8 Distributed Set get - PASS
• 17.9 Distributed Set size - PASS
• 17.10 Distributed Set add() - PASS
• 17.11 Distributed Set addAll() - PASS
• 17.12 Distributed Set contains() - PASS
• 17.13 Distributed Set containsAll() - PASS
• 17.14 Distributed Set remove() - FAIL
• Set remove was incorrect
• 17.15 Distributed Set removeAll() - FAIL
• Set removeAll was incorrect
• 17.16 Distributed Set addAll() - FAIL
• Set addAll was incorrect
• 17.17 Distributed Set clear() - FAIL
• Set clear was incorrect
• 17.18 Distributed Set addAll() - PASS
• 17.19 Distributed Set retain() - PASS
• 17.20 Partitioned Transactional maps put - PASS
• 17.21 Partitioned Transactional maps get - PASS

Case 62: Healing Partition - PASS

• 62.1 Deleteing firewall rules - PASS

Case 8: Compare ONOS Topology view to Mininet topology - FAIL

Compare topology objects between Mininet and ONOS

• 8.1 Comparing ONOS topology to MN topology - FAIL
• ONOS topology don't match Mininet
• 8.2 Hosts view is consistent across all ONOS nodes - PASS
• 8.3 Hosts information is correct - PASS
• 8.4 Host attachment points to the network - PASS
• 8.5 Clusters view is consistent across all ONOS nodes - FAIL
• ONOS nodes have different views of clusters
• 8.6 There is only one SCC - FAIL
• ONOS shows 14 SCCs
• 8.7 Device information is correct - FAIL
• Device information is incorrect
• 8.8 Links are correct - FAIL
• Links are incorrect
• 8.9 Hosts are correct - PASS
• 8.10 Checking ONOS nodes - PASS

Case 7: Running ONOS Constant State Tests - FAIL

• 7.1 Check that each switch has a master - FAIL
• Some devices don't have a master assigned
• 7.2 Read device roles from ONOS - PASS
• 7.3 Check for consistency in roles from each controller - FAIL
• ONOS nodes have different views of switch roles
• 7.4 Get the intents and compare across all nodes - PASS
• 7.5 Check for consistency in Intents from each controller - PASS
• 7.6 Compare current intents with intents before the failure - PASS
• 7.7 Get the OF Table entries and compare to before component failure - FAIL
• Changes were found in the flow tables
• 7.8 Leadership Election is still functional - FAIL
• Something went wrong with Leadership election

Case 4: Verify connectivity by sending traffic across Intents - FAIL

Ping across added host intents to check functionality and check the state of the intent

• 4.1 Check Intent state - PASS
• 4.2 Ping across added host intents - FAIL
• Intents have not been installed correctly, pings failed.
• 4.3 Check leadership of topics - PASS
• 4.4 Wait a minute then ping again - FAIL
• Intents have not been installed correctly, pings failed.

Case 15: Check that Leadership Election is still functional - FAIL

• 15.1 Run for election on each node - PASS
• 15.2 Check that each node shows the same leader and candidates - FAIL
• Nodes have different leaderboards
• 15.3 Find current leader and withdraw - FAIL
• Node was not withdrawn from election
• 15.4 Check that a new node was elected leader - FAIL
• Something went wrong with Leadership election
• 15.5 Check that that new leader was the candidate of old leader - FAIL
• Incorrect Candidate Elected
• 15.6 Run for election on old leader( just so everyone is in the hat ) - FAIL
• App failed to run for election
• 15.7 Check that oldLeader is a candidate, and leader if only 1 node - FAIL
• Something went wrong with Leadership election after the old leader re-ran for election