You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Ivan Rakov (Jira)" <ji...@apache.org> on 2020/08/17 17:09:00 UTC
[jira] [Commented] (IGNITE-13253) Advanced heuristics for
historical rebalance
[ https://issues.apache.org/jira/browse/IGNITE-13253?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17179132#comment-17179132 ]
Ivan Rakov commented on IGNITE-13253:
-------------------------------------
Dev list discussion: http://apache-ignite-developers.2346864.n4.nabble.com/Choosing-historical-rebalance-heuristics-td48389.html#a48409
> Advanced heuristics for historical rebalance
> --------------------------------------------
>
> Key: IGNITE-13253
> URL: https://issues.apache.org/jira/browse/IGNITE-13253
> Project: Ignite
> Issue Type: Improvement
> Reporter: Vladislav Pyatkov
> Assignee: Ivan Rakov
> Priority: Major
>
> Before, cluster detects partitions that have not to rebalance by history, by them size. This threshold might be set through a system property IGNITE_PDS_WAL_REBALANCE_THRESHOLD. But it is not fair deciding which partitions will be rebalanced by WAL only by them size. WAL can have much more records than size of a partition (many update by one key) and that rebalance required more data than full transferring by network.
> Need to implement a heuristic, that might to estimate data size.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)