TopologyManager's nodesOnHost map gets corrupted after several connection/disconnection #2384 #2386
Conversation
synchronize fork
…connection/disconnection - replaced LinkedList with LinkedHashSet (avoids adding the same Node several time) - added defensive checks.
| @@ -19,4 +19,5 @@ pa.rm.select.node.dynamicity=10000 | |||
| pa.rm.node.source.ping.frequency=10000 | |||
| pa.rm.client.ping.frequency=10000 | |||
| pa.rm.select.script.timeout=10000 | |||
| pa.rm.topology.enabled=false | |||
There was a problem hiding this comment.
Why is it now required to enable topology in all configuration files even when it was not before? Is it because of the change in test LocalSelectionTest?
There was a problem hiding this comment.
The LocalSelectionTest does need the topology to be enabled. The JobLegacySchema test in the scheduler needs both the topology and the topology-distance in order to pass.
|
Is it not possible to write a test to reproduce the issue? if it is a concurrency issue, defensive checks will not be enough. |
|
Right now I am not able to reproduce the issue, I added some debug information to try to understand it a bit more. Then I will be able to write a test eventually. |
| } | ||
| result.getExtraNodes().addAll(newExtraNodes); | ||
| } | ||
| if (--number <= 0) { |
There was a problem hiding this comment.
Refactor this code to not nest more than 3 if/for/while/switch/try statements. 
|
SonarQube analysis reported 5 issues:
Watch the comments in this conversation to review them.
|
| @@ -245,6 +264,38 @@ public Topology getTopology() { | |||
| } | |||
| } | |||
|
|
|||
| private NodeSet getNodeSetWithExtraNodes(Set<Node> nodes, int number) { | |||
There was a problem hiding this comment.
Could you give better naming to the number variable?
| public class TestTopologyConcurrency { | ||
|
|
||
| // number of thread used in this test | ||
| final int nb_threads = 20; |
There was a problem hiding this comment.
Java code convention is Camel case.
There was a problem hiding this comment.
"Test" should be always at the end of the name of the test class and at the beginning of the test methods :)
TopologyManager's nodesOnHost map gets corrupted after several connection/disconnection #2384
No description provided.