You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Kirill Tkalenko (Jira)" <ji...@apache.org> on 2023/06/08 14:19:00 UTC
[jira] [Commented] (IGNITE-19685) Treat Scalecube's LEAVING event as 'node left' in the physical topology
[ https://issues.apache.org/jira/browse/IGNITE-19685?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17730584#comment-17730584 ]
Kirill Tkalenko commented on IGNITE-19685:
------------------------------------------
Looks good.
> Treat Scalecube's LEAVING event as 'node left' in the physical topology
> -----------------------------------------------------------------------
>
> Key: IGNITE-19685
> URL: https://issues.apache.org/jira/browse/IGNITE-19685
> Project: Ignite
> Issue Type: Improvement
> Reporter: Roman Puchkovskiy
> Assignee: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-beta2
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> When a Scalecube's Cluster instance shuts down, the local member tells everyone that it is LEAVING (it's the only reason a member might end up in this state). Other nodes do not treat the member as REMOVED immediately as they see it LEAVING; instead, a whole suspicion interval must pass before the member is considered to be REMOVED.
> As we know that the LEAVING member has already left (and will not be back with the same ID), we should treat the LEAVING event as 'Node left' (in the physical topology).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)