Mailing-List: contact dev-help@ignite.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@ignite.apache.org
Date: Tue, 14 Nov 2017 13:57:01 +0000 (UTC)
From: "Vladimir Ozerov (JIRA)" <jira@apache.org>
To: dev@ignite.apache.org
Message-ID: <JIRA.13118255.1510667804000.232910.1510667821530@Atlassian.JIRA>
In-Reply-To: <JIRA.13118255.1510667804000@Atlassian.JIRA>
References: <JIRA.13118255.1510667804000@Atlassian.JIRA> <JIRA.13118255.1510667804562@jira-lw-us.apache.org>
Subject: [jira] [Created] (IGNITE-6904) SQL: Partition reservations are
 released too early in lazy mode
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit
archived-at: Tue, 14 Nov 2017 13:57:07 -0000

Vladimir Ozerov created IGNITE-6904:
---------------------------------------

             Summary: SQL: Partition reservations are released too early in lazy mode
                 Key: IGNITE-6904
                 URL: https://issues.apache.org/jira/browse/IGNITE-6904
             Project: Ignite
          Issue Type: Task
      Security Level: Public (Viewable by anyone)
          Components: sql
    Affects Versions: 2.3
            Reporter: Vladimir Ozerov
             Fix For: 2.4


In lazy mode we advance query execution as new page requests arrive. However, method {{GridMapQueryExecutor#onQueryRequest0}} releases partition reservations when only the very first page is processed:
{code}
finally {
    GridH2QueryContext.clearThreadLocal();

    if (distributedJoinMode == OFF)
        qctx.clearContext(false);
}
{code}

It means that incorrect results may be returned on unstable topology. We need to release partitions only after the whole query is executed.


--
This message was sent by Atlassian JIRA
(v6.4.14#64029)