You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chesnay Schepler (Jira)" <ji...@apache.org> on 2020/12/13 22:46:00 UTC
[jira] [Closed] (FLINK-19490) Document SOMAXCONN OS limit on blob
server backlog
[ https://issues.apache.org/jira/browse/FLINK-19490?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chesnay Schepler closed FLINK-19490.
------------------------------------
Resolution: Fixed
master: 8b05cbee4425c5ee33d73bed1473e075d7e17387
> Document SOMAXCONN OS limit on blob server backlog
> --------------------------------------------------
>
> Key: FLINK-19490
> URL: https://issues.apache.org/jira/browse/FLINK-19490
> Project: Flink
> Issue Type: Improvement
> Components: Documentation, Runtime / Coordination
> Reporter: Chesnay Schepler
> Assignee: Chesnay Schepler
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.13.0
>
>
> The {{BlobServer}} has a {{blob.fetch.backlog}} setting that controls how many connections can be queued when {{blob.fetch.num-concurrent}} is exceeded.
> However, the backlog relies on OS functionality, which has it's own maximum limit (SOMAXCONN). Put another way, our backlog setting is just a hint to the OS.
> This should be documented so that users who are thinking about this option are aware of it's limitations.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)