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

Posted to issues@ignite.apache.org by "Mirza Aliev (Jira)" <ji...@apache.org> on 2022/05/19 15:45:00 UTC

[jira] [Updated] (IGNITE-17013) Implement failure handling for changePeerAsync invocation regarding rebalance algorithm

     [ https://issues.apache.org/jira/browse/IGNITE-17013?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Mirza Aliev updated IGNITE-17013:
---------------------------------
    Description: According to the relalance algorithm https://github.com/apache/ignite-3/blob/main/modules/table/tech-notes/rebalance.md after leader node receives an updating of a pending key, it starts a raft group reconfiguration by calling {{RaftGroupService#changePeersAsync}} which returns future, that can be completed exceptionally. We should provide the way to retry {{RaftGroupService#changePeersAsync}} invocation, with some throttling, and in case of failure cancel the current rebalance.     (was: According to the relalance algorithm https://github.com/apache/ignite-3/blob/main/modules/table/tech-notes/rebalance.md after leader node receives an updating of a pending key, it starts raft group reconfiguration by calling {{RaftGroupService#changePeersAsync}} which returns future, that can be completed exceptionally. We should provide the way to retry {{RaftGroupService#changePeersAsync}} invocation, with some throttling, and in case of failure cancel the current rebalance.   )

> Implement failure handling for changePeerAsync invocation regarding rebalance algorithm  
> -----------------------------------------------------------------------------------------
>
>                 Key: IGNITE-17013
>                 URL: https://issues.apache.org/jira/browse/IGNITE-17013
>             Project: Ignite
>          Issue Type: Task
>            Reporter: Mirza Aliev
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.0.0-alpha5
>
>
> According to the relalance algorithm https://github.com/apache/ignite-3/blob/main/modules/table/tech-notes/rebalance.md after leader node receives an updating of a pending key, it starts a raft group reconfiguration by calling {{RaftGroupService#changePeersAsync}} which returns future, that can be completed exceptionally. We should provide the way to retry {{RaftGroupService#changePeersAsync}} invocation, with some throttling, and in case of failure cancel the current rebalance.   



--
This message was sent by Atlassian Jira
(v8.20.7#820007)