[ https://issues.apache.org/jira/browse/FLINK-33532?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17786645#comment-17786645 ]
Yangze Guo commented on FLINK-33532: ------------------------------------ master: 8ef71ba1f190a5cf6093adb6c40c6a6d04944d20 d18a4bfe596fc580f8280750fa3bfa22007671d9 > Move the serialization of ShuffleDescriptorGroup out of the RPC main thread > --------------------------------------------------------------------------- > > Key: FLINK-33532 > URL: https://issues.apache.org/jira/browse/FLINK-33532 > Project: Flink > Issue Type: Sub-task > Components: Runtime / Network > Reporter: Yangze Guo > Assignee: dizhou cao > Priority: Major > Labels: pull-request-available > Fix For: 1.19.0 > > > Currently, the serilization of ShuffleDescriptorGroup would happen in the RPC > main thread even for light jobs. That would harm the QPS of OLAP session, > e.g. the deserialization will take 3ms at TM side for each task of a 128*128 > wordcount job. We propose to move it out of the rpc main thread. For heavy > jobs, it still worth to cache the serialized value for memory efficiency. At > the same time we need to make the serialization thread pool configurable. -- This message was sent by Atlassian Jira (v8.20.10#820010)