[ 
https://issues.apache.org/jira/browse/IGNITE-6219?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16147140#comment-16147140
 ] 

ASF GitHub Bot commented on IGNITE-6219:
----------------------------------------

GitHub user dkarachentsev opened a pull request:

    https://github.com/apache/ignite/pull/2548

    IGNITE-6219 - IgniteCache#loadCache executes local load in caller thread

    

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/gridgain/apache-ignite ignite-6219

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/ignite/pull/2548.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #2548
    
----
commit fcda4f1af212883f2b12f3185a5f897157f82a8d
Author: dkarachentsev <[email protected]>
Date:   2017-08-30T10:39:05Z

    IGNITE-6219 - IgniteCache#loadCache executes local load in caller thread

----


> IgniteCache#loadCache executes local load in caller thread
> ----------------------------------------------------------
>
>                 Key: IGNITE-6219
>                 URL: https://issues.apache.org/jira/browse/IGNITE-6219
>             Project: Ignite
>          Issue Type: Bug
>          Components: cache
>    Affects Versions: 2.1
>            Reporter: Valentin Kulichenko
>            Assignee: Dmitry Karachentsev
>            Priority: Critical
>             Fix For: 2.2
>
>
> {{IgniteCache#loadCache}} method broadcasts an internal task under the hood. 
> If one of the jobs are local (i.e. if {{loadCache}} is invoked on server 
> node), this job is executed in a caller thread, potentially *before all or 
> some remote requests are sent*. Since data loading is generally long running 
> process, its duration doubles in this scenario.
> Possible solution is to check the list of nodes before task execution, and if 
> local node is there, execute on remote nodes first, and only then submit to 
> local node. This way we make sure that remote nodes never wait for the local 
> node.



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

Reply via email to