[
https://issues.apache.org/jira/browse/CASSANALYTICS-193?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yifan Cai updated CASSANALYTICS-193:
------------------------------------
Labels: lhf (was: )
> Sidecar Client maxPoolSize Option Silently Ignored by HTTP Connection Pool
> --------------------------------------------------------------------------
>
> Key: CASSANALYTICS-193
> URL: https://issues.apache.org/jira/browse/CASSANALYTICS-193
> Project: Apache Cassandra Analytics
> Issue Type: Bug
> Components: Reader
> Reporter: Yifan Cai
> Priority: Normal
> Labels: lhf
>
> The Sidecar client's maxPoolSize configuration option is only applied to
> Vert.x's internal worker thread pool (Sidecar.java), never to the underlying
> HttpClientConfig used by VertxHttpClient to build the WebClient. As a result,
> the HTTP connection pool that actually bounds concurrent SSTable-download
> requests silently falls back to HttpClientConfig's own hardcoded default of
> 20 connections, regardless of what a user configures. This can cause download
> requests to queue behind the connection limit and stall long enough to trip
> the 600s BufferingInputStream timeout during bulk reads.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
