[
https://issues.apache.org/jira/browse/IGNITE-19747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vladislav Pyatkov updated IGNITE-19747:
---------------------------------------
Description:
The test have suspicious history on TC:
[https://ci.ignite.apache.org/test/3244623774880880898?currentProjectId=ApacheIgnite3xGradle_Test_IntegrationTests]
In all cases, it is an exception like this:
{noformat}
org.apache.ignite.lang.IgniteInternalException: IGN-CMN-65535
TraceId:fcdca6b2-e939-4e79-b03f-79ac4dca5d16 No such partition 0 in table TBL1
at
app//org.apache.ignite.internal.table.distributed.storage.InternalTableImpl.partitionRaftGroupService(InternalTableImpl.java:1141)
at
app//org.apache.ignite.internal.configuration.storage.ItRebalanceDistributedTest.testOnLeaderElectedRebalanceRestart(ItRebalanceDistributedTest.java:386)
at
java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)
at
java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
{noformat}
The root cause is we have no guaranty that a table available on each node if we
got it only at one.
{code:title=Test code fragment}
TableImpl table = (TableImpl) await(nodes.get(1).tableManager.createTableAsync(
"TBL1",
zoneName,
tblChanger -> SchemaConfigurationConverter.convert(schTbl1,
tblChanger)));
...
TableImpl nonLeaderTable = (TableImpl)
findNodeByConsistentId(nonLeaderNodeConsistentId).tableManager.table("TBL1");
...
nonLeaderTable.internalTable().partitionRaftGroupService(0)
{code}
We wait the table on node with index 1, but expected that is completed on
another node (nonLeaderNodeConsistentId).
was:
The test have suspicious history on TC:
[https://ci.ignite.apache.org/test/3244623774880880898?currentProjectId=ApacheIgnite3xGradle_Test_IntegrationTests]
In all cases, it is an exception like this:
{noformat}
org.apache.ignite.lang.IgniteInternalException: IGN-CMN-65535
TraceId:fcdca6b2-e939-4e79-b03f-79ac4dca5d16 No such partition 0 in table TBL1
at
app//org.apache.ignite.internal.table.distributed.storage.InternalTableImpl.partitionRaftGroupService(InternalTableImpl.java:1141)
at
app//org.apache.ignite.internal.configuration.storage.ItRebalanceDistributedTest.testOnLeaderElectedRebalanceRestart(ItRebalanceDistributedTest.java:386)
at
java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)
at
java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
{noformat}
> ItRebalanceDistributedTest.testOnLeaderElectedRebalanceRestart is flacky
> ------------------------------------------------------------------------
>
> Key: IGNITE-19747
> URL: https://issues.apache.org/jira/browse/IGNITE-19747
> Project: Ignite
> Issue Type: Bug
> Reporter: Vladislav Pyatkov
> Priority: Major
> Labels: ignite-3
>
> The test have suspicious history on TC:
> [https://ci.ignite.apache.org/test/3244623774880880898?currentProjectId=ApacheIgnite3xGradle_Test_IntegrationTests]
> In all cases, it is an exception like this:
> {noformat}
> org.apache.ignite.lang.IgniteInternalException: IGN-CMN-65535
> TraceId:fcdca6b2-e939-4e79-b03f-79ac4dca5d16 No such partition 0 in table TBL1
> at
> app//org.apache.ignite.internal.table.distributed.storage.InternalTableImpl.partitionRaftGroupService(InternalTableImpl.java:1141)
> at
> app//org.apache.ignite.internal.configuration.storage.ItRebalanceDistributedTest.testOnLeaderElectedRebalanceRestart(ItRebalanceDistributedTest.java:386)
> at
> java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
> Method)
> at
> java.base@11.0.17/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
> {noformat}
> The root cause is we have no guaranty that a table available on each node if
> we got it only at one.
> {code:title=Test code fragment}
> TableImpl table = (TableImpl)
> await(nodes.get(1).tableManager.createTableAsync(
> "TBL1",
> zoneName,
> tblChanger -> SchemaConfigurationConverter.convert(schTbl1,
> tblChanger)));
> ...
> TableImpl nonLeaderTable = (TableImpl)
> findNodeByConsistentId(nonLeaderNodeConsistentId).tableManager.table("TBL1");
> ...
> nonLeaderTable.internalTable().partitionRaftGroupService(0)
> {code}
> We wait the table on node with index 1, but expected that is completed on
> another node (nonLeaderNodeConsistentId).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
