[
https://issues.apache.org/jira/browse/IGNITE-25262?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17948925#comment-17948925
]
Roman Puchkovskiy commented on IGNITE-25262:
--------------------------------------------
On a node that hosts primary replica of a partition, a write is done by replica
(PartitionReplicaListener) and also Raft state machine (PartitionListener). We
have protection against such double writes (second write is rejected in
PartitionListener). This protection relies on identifying the group using
TablePartitionId, so it doesn't work for colocation mode, where ZonePartitionId
is to be used.
The fix does just that.
> ItRwTransactionAndIndexesTest.testDropIndexInsideRwTransaction fails in case
> enabled colocation
> -----------------------------------------------------------------------------------------------
>
> Key: IGNITE-25262
> URL: https://issues.apache.org/jira/browse/IGNITE-25262
> Project: Ignite
> Issue Type: Bug
> Reporter: Alexander Lapin
> Assignee: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
> Time Spent: 10m
> Remaining Estimate: 0h
>
> {code:java}
> org.mockito.exceptions.verification.TooManyActualInvocations:
> testHashIndexStorage.put(<any>);
> Wanted 1 time:
> -> at
> org.apache.ignite.internal.storage.index.impl.TestHashIndexStorage.put(TestHashIndexStorage.java:64)
> But was 2 times:
> -> at
> org.apache.ignite.internal.table.distributed.TableSchemaAwareIndexStorage.put(TableSchemaAwareIndexStorage.java:83)
> -> at
> org.apache.ignite.internal.table.distributed.TableSchemaAwareIndexStorage.put(TableSchemaAwareIndexStorage.java:83)
> {code}
> Reproduced locally.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
