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 2021/01/18 12:32:00 UTC

[jira] [Commented] (IGNITE-13994) Rebalance huge cache in the case of in-memory cluster

    [ https://issues.apache.org/jira/browse/IGNITE-13994?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17267215#comment-17267215 ] 

Anton Vinogradov commented on IGNITE-13994:
-------------------------------------------

We should test both cases, a huge cache, and a huge entry.

> Rebalance huge cache in the case of in-memory cluster
> -----------------------------------------------------
>
>                 Key: IGNITE-13994
>                 URL: https://issues.apache.org/jira/browse/IGNITE-13994
>             Project: Ignite
>          Issue Type: Test
>            Reporter: Ivan Daschinskiy
>            Priority: Major
>              Labels: IEP-56
>
> There are some evidence, that rebalancing huge cache without rebalance throttling can cause OOM on supplier. We need to cover this scenario.
> Scenario:
> 1. Start two nodes and 1 replicated cache with data region much more than heap.
> 2. Stop one of the node.
> 3. Load data to cache almost equal to size of data region.
> 4. Start node.
> Goal is to run experiments with parameters
> 1. Heap size
> 2. Cache size
> 3. Rebalance batch size.
> 4. Rebalance throttle



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