[jira] [Commented] (IGNITE-1698) SqlFieldsQuery works incorrectly in case of topology changes

["Sergi Vladykin (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-1698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14961952#comment-14961952 ] 

Sergi Vladykin commented on IGNITE-1698:
----------------------------------------

I've merged the pull request with the test case into master (with few minor changes).

The problem is that if we have 0 backups we don't create filter for keys by primary partition (we assume that we already have only primary entries and thus there is nothing to filter). But actually we've successfully reserved primary partitions but few partitions which used to be primary on our node from previous topology version were not evicted yet and we are getting extra entries from these ex-primary partitions. We need to add check into IgniteH2Indexing#backupFilter that eviction of ex-primary partitions is finished and we will see only primary partitions.



> SqlFieldsQuery works incorrectly in case of topology changes
> ------------------------------------------------------------
>
>                 Key: IGNITE-1698
>                 URL: https://issues.apache.org/jira/browse/IGNITE-1698
>             Project: Ignite
>          Issue Type: Bug
>    Affects Versions: ignite-1.4
>            Reporter: Andrey Gura
>
> {{SqlFieldsQuery}} returns duplicate rows. Steps to reproduce:
> # Start one node
> # Create cache nad populate data.
> # Start second node
> # Execute query
> Result: {{QueryCursor}} retrieves duplicate rows.
> See corresponding test in PR https://github.com/apache/ignite/pull/162
> {{testSqlFieldsQuery}} passes, but {{testSqlFieldsQueryWithTopologyChanges}} fails.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)