[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-644640968 The failed case `StreamingKafkaITCase` is unrelated. The problem is FLINK-18311. Merging. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-644476328 @flinkbot run azure This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-643205521 I just found a problem that in `DefaultExecutionSlotAllocator` the slot request canceler was registered with an incorrect slot request id. It is fixed now and the diff can be found at https://github.com/zhuzhurk/flink/commit/45e5e7f306e5054ff0905879696accb6f1c282ef. @azagrebin would you take another look? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-643052813 @flinkbot run azure This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-643007617 @flinkbot run azure This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-635178001 @azagrebin @GJL #12375 is opened for bulk slot allocation part with a `BulkSlotProvider` introduced. Related comments above are also addressed in it. Would you take a look? This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
[GitHub] [flink] zhuzhurk commented on pull request #12256: [FLINK-17018][runtime] Allocates slots in bulks for pipelined region scheduling
zhuzhurk commented on pull request #12256: URL: https://github.com/apache/flink/pull/12256#issuecomment-633867170 Thanks for the comments! @azagrebin Agreed that we can make new allocator and bulk slot allocation independent from existing `SchedulerImp` and `DefaultExecutionSlotAllocator ` for further evolving. Will do it. >> Also, maybe we could split the PR somehow into e.g. two PRs: bulk slot allocation and OneSlotPerExecutionSlotAllocator to be able to merge smaller chunks of code separately Sure we can do it. The PR is already in smaller pieces and it would be easy to split the PR. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org