Aleksey Plekhanov created IGNITE-27883:
------------------------------------------
Summary: Flaky TcpCommunicationSpiDropNodesTest.testOneNode
Key: IGNITE-27883
URL: https://issues.apache.org/jira/browse/IGNITE-27883
Project: Ignite
Issue Type: Bug
Reporter: Aleksey Plekhanov
Assignee: Aleksey Plekhanov
Throws an error on TeamCity:
{noformat}
java.lang.AssertionError: Should have exception here.
at org.junit.Assert.fail(Assert.java:88)
at
org.apache.ignite.testframework.junits.JUnitAssertAware.fail(JUnitAssertAware.java:50)
at
org.apache.ignite.spi.communication.tcp.TcpCommunicationSpiDropNodesTest.testOneNode(TcpCommunicationSpiDropNodesTest.java:138)
{noformat}
Can be reproduced locally by changing sleep time in test to 1500 for example.
Reason:
When second node join the grid, partiition resend is scheduled (see
GridCachePartitionExchangeManager#scheduleResendPartitions) with timeout 1500
ms (see GridCachePartitionExchangeManager#partResendTimeout). Partition resend
opens TCP communication connection from node0 to all other nodes. If it happens
before TCP communication for node2 blocked, connection can be established and
survive idle timeout.
It happens when time between second node join and TCP communication blocking is
about 1500 ms (including current 1000 ms sleep time).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
