[jira] [Commented] (IGNITE-12181) Rebalance hangs on BLT change on cluster with in-memory regions

["Maxim Muzafarov (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-12181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16941793#comment-16941793 ] 

Maxim Muzafarov commented on IGNITE-12181:
------------------------------------------

It seems to me that the issue is related to these lines:

{code:title=RowStore.java:135}
    public void addRows(Collection<? extends CacheDataRow> rows,
        IoStatisticsHolder statHolder) throws IgniteCheckedException {
        assert ctx.database().checkpointLockIsHeldByThread(); <- HERE

        freeList.insertDataRows(rows, statHolder);
    }
{code}

{code:title=GridDhtPartitionDemander.java:981}
    private boolean preloadEntry(CacheDataRow row, AffinityTopologyVersion topVer) throws IgniteCheckedException {
        assert !grp.mvccEnabled();
        assert ctx.database().checkpointLockIsHeldByThread();

        ...
    }
{code}



> Rebalance hangs on BLT change on cluster with in-memory regions
> ---------------------------------------------------------------
>
>                 Key: IGNITE-12181
>                 URL: https://issues.apache.org/jira/browse/IGNITE-12181
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Maxim Muzafarov
>            Assignee: Maxim Muzafarov
>            Priority: Blocker
>             Fix For: 2.8
>
>         Attachments: RebalanceInMemoryWithPersistence.java
>
>
> # Configured BLT on the _node-1_ (cluster activated)
> # Configured in-memory region with a single cache on the _node-1_
> # A new _node-2_  join to the cluster
> # Set new BLT based on two nodes
> # Rebalance hangs
> Reproducer attached.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)