You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@ignite.apache.org by "Anton Vinogradov (JIRA)" <ji...@apache.org> on 2018/10/02 10:27:00 UTC

[jira] [Commented] (IGNITE-9491) Exchange latch coordinator shouldn't be oldest node in a cluster

    [ https://issues.apache.org/jira/browse/IGNITE-9491?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16635259#comment-16635259 ] 

Anton Vinogradov commented on IGNITE-9491:
------------------------------------------

The better case is to have special coordinator nodes. 
For example - the oldest non-affinity node (empty).

> Exchange latch coordinator shouldn't be oldest node in a cluster
> ----------------------------------------------------------------
>
>                 Key: IGNITE-9491
>                 URL: https://issues.apache.org/jira/browse/IGNITE-9491
>             Project: Ignite
>          Issue Type: Improvement
>          Components: cache
>    Affects Versions: 2.5
>            Reporter: Pavel Kovalenko
>            Assignee: Pavel Kovalenko
>            Priority: Major
>             Fix For: 2.8
>
>
> Currently, we have a lot of components having coordinator election ability. Each of these components is electing the oldest node as coordinator. It leads to overloading the oldest node and may be a cause of delaying of some processes.
> The oldest node can have large inbound/outbound messages queue in large topologies which leads to delaying of processing Exchange Latch Ack messages. We should choose secondary oldest node as coordinator to unload the oldest coordinator. This change will significantly accelerate exchange latch waiting process.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)