You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Steve Loughran (Jira)" <ji...@apache.org> on 2025/03/04 11:46:00 UTC
[jira] [Commented] (HADOOP-19478) S3A: pull out new configuration load/probes under S3AStore
[ https://issues.apache.org/jira/browse/HADOOP-19478?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17932288#comment-17932288 ]
Steve Loughran commented on HADOOP-19478:
-----------------------------------------
update: not going to do it in the conditional create...to complex and it will delay merging of that.
> S3A: pull out new configuration load/probes under S3AStore
> ----------------------------------------------------------
>
> Key: HADOOP-19478
> URL: https://issues.apache.org/jira/browse/HADOOP-19478
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Affects Versions: 3.4.1
> Reporter: Steve Loughran
> Assignee: Steve Loughran
> Priority: Major
>
> I am doing this as part of HADOOP-19256, but wanted to call it out for specific mention.
> S3AFS i's too full of switches; it is why initialize() is so big, and there are lots and lots of fields to record the values.
> Many of these need to go down into S3AStoreImpl, but replicating the same design just pushes the mess down.
> Proposed: a child service StoreConfigurationService, which reads in the config during serviceInit(), and set the state in there, from where it can be probed.
> I am initially doing this purely for new configuration flags for the conditional write feature
> * moving other flags in there would be separate work
> * new boolean config options should go in here
> * we also need to think about integers, units of scale and durations.
> Ideally, we should use something reflection based as abfs does.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org