[
https://issues.apache.org/jira/browse/HADOOP-19672?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18032070#comment-18032070
]
ASF GitHub Bot commented on HADOOP-19672:
-----------------------------------------
bhattmanish98 commented on PR #7967:
URL: https://github.com/apache/hadoop/pull/7967#issuecomment-3432065456
@steveloughran We’re performing cache refresh when the number of available
connections in the cache drops below a certain threshold. The warmup normally
happens during file system initialization, but if the file system is already
active and making multiple parallel network calls, creating new connections on
demand with the Apache client tends to take longer.
To avoid delays in such cases, we asynchronously pre-create and cache a few
additional connections whenever the pool runs low. This ensures that subsequent
network calls can use already-established connections and spend time on actual
data processing rather than connection setup.
> ABFS: Network Error-Based Client Switchover: Apache to JDK (continuous
> failure))
> --------------------------------------------------------------------------------
>
> Key: HADOOP-19672
> URL: https://issues.apache.org/jira/browse/HADOOP-19672
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/azure
> Affects Versions: 3.4.2
> Reporter: Manish Bhatt
> Assignee: Manish Bhatt
> Priority: Major
> Labels: pull-request-available
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
