Hi, what do expect the performance gain to be by using volcano versus standard scheduler.
Just to be sure there are two aspects here. 1. Procuring the Kubernetes cluster 2. Running the job through spark-submit Item 1 is left untouched and we should see improvements in item 2 with Volcano Thanks view my Linkedin profile <https://www.linkedin.com/in/mich-talebzadeh-ph-d-5205b2/> https://en.everybodywiki.com/Mich_Talebzadeh *Disclaimer:* Use it at your own risk. Any and all responsibility for any loss, damage or destruction of data or any other property which may arise from relying on this email's technical content is explicitly disclaimed. The author will in no case be liable for any monetary damages arising from such loss, damage or destruction. On Thu, 24 Feb 2022 at 03:35, Yikun Jiang <yikunk...@gmail.com> wrote: > First, much thanks for all your help (Spark/Volcano/Yunikorn community) to > make this SPIP happen! > > Especially,@dongjoon-hyun @holdenk @william-wang @attilapiros @HyukjinKwon > @martin-g @yangwwei @tgravescs > > The SPIP is near the end of the stage. It can be said that it is beta > available at the basic level. > > I also draft a simple slide to show how to use and help you understand > what we have done: > > https://docs.google.com/presentation/d/1XDsTWPcsBe4PQ-1MlBwd9pRl8mySdziE_dJE6iATNw8 > > Below are also some recap to help you understand current implementation > and next step on SPIP: > > *# Existing work* > *## Basic part:* > - SPARK-36059 <https://issues.apache.org/jira/browse/SPARK-36059> *New > configuration:* ability to specify "schedulerName" in driver/executor for > Spark on K8S > - SPARK-37331 <https://issues.apache.org/jira/browse/SPARK-37331> *New > workflow：*ability to create pre-populated resources before driver pod > for Spark on K8S > - SPARK-37145 <https://issues.apache.org/jira/browse/SPARK-37145> *New > developer API:* support user feature step with configuration for Spark on > K8S > - *(reviewing)* *New Job Configurations* for Spark on K8S: > - SPARK-38188 <https://issues.apache.org/jira/browse/SPARK-38188>: > spark.kubernetes.job.queue > - SPARK-38187 <https://issues.apache.org/jira/browse/SPARK-38187>: > spark.kubernetes.job.[minCPU|minMemory] > - SPARK-38189 <https://issues.apache.org/jira/browse/SPARK-38189>: > spark.kubernetes.job.priorityClassName > > *## Volcano Part:* > - SPARK-37258 <https://issues.apache.org/jira/browse/SPARK-37258> *New > volcano extension* in kubernetes-client fabric8io/kubernetes-client#3579 > - SPARK-36061 <https://issues.apache.org/jira/browse/SPARK-36061> *New > profile: *-Pvolcano > - SPARK-36061 <https://issues.apache.org/jira/browse/SPARK-36061> *New > Feature Step:* VolcanoFeatureStep > - SPARK-36061 <https://issues.apache.org/jira/browse/SPARK-36061> *New > integration test:* > *- Passed on x86 and Arm64 (Linux on Huawei Kunpeng 920 and MacOS on > Apple Silicon M1).* > - Test basic volcano workflow > - Test all existing tests based on the volcano. > > *## Yunikorn Part:* > @yangwwei will also make the efforts for Yunikorn module feature step > since this week. > I will help to complete the yunikorn integration based on previous > experience. > > *# Next Plan* > There are also 3 main tasks to be completed before v3.3 code freeze: > 1. (reviewing) SPARK-38188 > <https://issues.apache.org/jira/browse/SPARK-38188>: Support queue > scheduling configuration > https://github.com/apache/spark/pull/35553 > 2. (reviewing) SPARK-38187 > <https://issues.apache.org/jira/browse/SPARK-38187>: Support resource > reservation (minCPU/minMemory configuration) > https://github.com/apache/spark/pull/35640 > 3. (reviewing) SPARK-38187 > <https://issues.apache.org/jira/browse/SPARK-38187>: Support priority > scheduling (priorityClass configuration): > https://issues.apache.org/jira/browse/SPARK-38189 > https://github.com/apache/spark/pull/35639 > 4. (WIP) SPARK-37809 <https://issues.apache.org/jira/browse/SPARK-37809>: > Yunikorn integration > > Also several misc work is gonna be completed before 3.3: > 1. Integrated volcano deploy into integration test (x86 and arm) > - Add it to spark kubernetes integration test once cross compile support: > https://github.com/volcano-sh/volcano/pull/1571 > 2. Complete doc and test guideline. > > Please feel free to contact me if you have any other concerns! Thanks! > > [1] https://issues.apache.org/jira/browse/SPARK-36057 >
