[jira] [Comment Edited] (IGNITE-18128) ReplicaIsAlreadyStartedException when starting a node having a table with at least 2 replicas

["Roman Puchkovskiy (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-18128?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17632624#comment-17632624 ] 

Roman Puchkovskiy edited comment on IGNITE-18128 at 11/12/22 8:23 AM:
----------------------------------------------------------------------

{{If the following modification is made in {{PartitionListener}}}}


{{private void handleSafeTimeSyncCommand(SafeTimeSyncCommand cmd, long commandIndex) {}}
{{    if (commandIndex <= storage.lastAppliedIndex()) }}{{{}}

{{        return;}}

{{    }}}

 

{{    storage.runConsistently(() -> {}}
{{{}        storage.lastAppliedIndex(commandIndex);{}}}{{{}return null;{}}}
{{    });}}
{{}}}


then the bug does not reproduce anymore.


was (Author: rpuch):
If the following modification is made in {{PartitionListener}}
private void handleSafeTimeSyncCommand(SafeTimeSyncCommand cmd, long commandIndex) \{
    // No-op.

    if (commandIndex <= storage.lastAppliedIndex()) {
        return;
    }

    storage.runConsistently(() -> \{
        storage.lastAppliedIndex(commandIndex);

        return null;
    });
}
then the bug does not reproduce anymore.

> ReplicaIsAlreadyStartedException when starting a node having a table with at least 2 replicas
> ---------------------------------------------------------------------------------------------
>
>                 Key: IGNITE-18128
>                 URL: https://issues.apache.org/jira/browse/IGNITE-18128
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Roman Puchkovskiy
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.0.0-beta2
>
>
> How to reproduce:
>  # Start a cluster of at least 2 nodes (I had 3)
>  # Create a table with at least 2 replicas (I had 3), wait for a couple of seconds so it gets replicated
>  # Restart one of the nodes
> {{ReplicaIsAlreadyStartedException}} is thrown because {{ReplicaManager#startReplica()}} is invoked twice for same replica.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)