You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Weijie Guo (Jira)" <ji...@apache.org> on 2023/04/18 17:49:00 UTC
[jira] [Comment Edited] (FLINK-25806) Remove legacy high availability services
[ https://issues.apache.org/jira/browse/FLINK-25806?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17713699#comment-17713699 ]
Weijie Guo edited comment on FLINK-25806 at 4/18/23 5:48 PM:
-------------------------------------------------------------
It seems that {{KubernetesMultipleComponentLeaderElectionHaServicesFactory}} is unused in the codebase and its code is exactly the same as {{KubernetesHaServicesFactory}}. I'm a bit curious if we can remove it? More aggressive, we can even remove both {{HaServicesFactory}} related to kubernetes and adopt a similar approach to {{HighAvailabilityServicesUtils#createZooKeeperHaServices}}.
One possible concern is that users may configure this in the following way:
{{high-availability.type=org.apache.flink.kubernetes.highavailability.KubernetesHaServicesFactory}}.
Especially before FLINK-27808, user can only enable k8s based HA service through this approach. I'm not sure how long we need to keep this compatibility. [~mapohl] and [~chesnay] Looking forward to hearing your thoughts.
was (Author: weijie guo):
It seems that {{KubernetesMultipleComponentLeaderElectionHaServicesFactory}} is unused in the codebase and its code is exactly the same as {{KubernetesHaServicesFactory}}. I'm a bit curious if we can remove it? More aggressive, we can even remove both {{HaServicesFactory}} related to kubernetes and adopt a similar approach to {{HighAvailabilityServicesUtils # createZooKeeperHaServices}}.
One possible concern is that users may configure this in the following way:
{{high-availability.type=org.apache.flink.kubernetes.highavailability.KubernetesHaServicesFactory}}.
Especially before FLINK-27808, user can only enable k8s based HA service through this approach. I'm not sure how long we need to keep this compatibility. [~mapohl] and [~chesnay] Looking forward to hearing your thoughts.
> Remove legacy high availability services
> ----------------------------------------
>
> Key: FLINK-25806
> URL: https://issues.apache.org/jira/browse/FLINK-25806
> Project: Flink
> Issue Type: Technical Debt
> Components: Runtime / Coordination
> Affects Versions: 1.16.0
> Reporter: Till Rohrmann
> Assignee: Chesnay Schepler
> Priority: Blocker
> Labels: pull-request-available
> Fix For: 1.16.0
>
>
> After FLINK-24038, we should consider removing the legacy high availability services {{ZooKeeperHaServices}} and {{KubernetesHaServices}} since they are now subsumed by the multiple component leader election service that only uses a single leader election per component.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)