You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Maksim Timonin (Jira)" <ji...@apache.org> on 2021/08/05 10:42:00 UTC

[jira] [Updated] (IGNITE-14944) Limited CacheQuery should not fail on onNodeLeft if PageStream has enough data

     [ https://issues.apache.org/jira/browse/IGNITE-14944?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Maksim Timonin updated IGNITE-14944:
------------------------------------
    Description: 
Currently onNodeLeft checks that left node is still running query, if it is then we fail. But if query is limited then we can have enough data in page buffer. So we can ignore such fail.

 

Also idea for limit optimization for MergeSort algorithm - add to comparison of streams metric of probability of loading next page (try use as much local data to avoid loading new pages): local node, size of current page for iteration.

  was:Currently onNodeLeft checks that left node is still running query, if it is then we fail. But if query is limited then we can have enough data in page buffer. So we can ignore such fail.


> Limited CacheQuery should not fail on onNodeLeft if PageStream has enough data
> ------------------------------------------------------------------------------
>
>                 Key: IGNITE-14944
>                 URL: https://issues.apache.org/jira/browse/IGNITE-14944
>             Project: Ignite
>          Issue Type: New Feature
>            Reporter: Maksim Timonin
>            Assignee: Maksim Timonin
>            Priority: Major
>
> Currently onNodeLeft checks that left node is still running query, if it is then we fail. But if query is limited then we can have enough data in page buffer. So we can ignore such fail.
>  
> Also idea for limit optimization for MergeSort algorithm - add to comparison of streams metric of probability of loading next page (try use as much local data to avoid loading new pages): local node, size of current page for iteration.



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