[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-24 Thread Astralidea
Github user Astralidea commented on the issue: https://github.com/apache/spark/pull/15588 @jerryshao I agree waiting time waste the CPU time, and I have tested @lw-lin mentioned feasible solution is not work in my environment. OK, If there is no better solution or advise. I will

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-24 Thread jerryshao
Github user jerryshao commented on the issue: https://github.com/apache/spark/pull/15588 I think this fix cannot really handle this imbalance receiver allocation problem, also blindly waste the CPU time. What @lw-lin mentioned is a feasible solution to wait for executors to be

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-23 Thread Astralidea
Github user Astralidea commented on the issue: https://github.com/apache/spark/pull/15588 @lw-lin spark.scheduler.minRegisteredResourcesRatio does not work. The reason it may could be I use mesos and it run executor not through driver. but I still need to make sure it have

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-22 Thread Astralidea
Github user Astralidea commented on the issue: https://github.com/apache/spark/pull/15588 @lw-lin Thanks for you reply. In my private cluster running spark is a little different. (I start drivr & executor by myself) I had try maxRegisteredWaitingTime, but I had not try minRegi

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-22 Thread lw-lin
Github user lw-lin commented on the issue: https://github.com/apache/spark/pull/15588 Spark Streaming would do a very simple dummy job ensure that all slaves have registered before scheduling the `Receiver`s; please see https://github.com/apache/spark/blob/v2.0.0/streaming/src/main/sc

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-22 Thread Astralidea
Github user Astralidea commented on the issue: https://github.com/apache/spark/pull/15588 @srowen But in my cluster I tested 10 times. 9 times successed, 1 time failed. Why not necessary? receiver balance scheduler affect performance. If new executor delay add to driver. rece

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-22 Thread srowen
Github user srowen commented on the issue: https://github.com/apache/spark/pull/15588 I don't think it's necessarily true that you want to wait for all receivers to begin processing. This change won't work in any event. --- If your project is set up for it, you can reply to this emai

[GitHub] spark issue #15588: [SPARK-18039][Scheduler] fix bug maxRegisteredWaitingTim...

2016-10-21 Thread AmplabJenkins
Github user AmplabJenkins commented on the issue: https://github.com/apache/spark/pull/15588 Can one of the admins verify this patch? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feat