回复：Flink application down due to RpcTimeout exception

2018-09-12 Thread Zhijiang(wangzhijiang999)
. 2. You can increase the default value of rpc timeout parameter(akka.ask.timeout) to work around temporarily. Best, Zhijiang -- 发件人：徐涛 发送时间：2018年9月13日(星期四) 14:10 收件人：user 主　题：Flink application down due to RpcTimeout exception Hi

回复：InpoolUsage & InpoolBuffers inconsistence

2018-09-17 Thread Zhijiang(wangzhijiang999)
%, there may be a lot of buffers queued in front of barrier, so the checkpoint may expire as you said. Best, Zhijiang -- 发件人：aitozi 发送时间：2018年9月18日(星期二) 12:59 收件人：user 主　题：Re: InpoolUsage & InpoolBuffers inconsistence And my d

回复：回复：InpoolUsage & InpoolBuffers inconsistence

2018-09-18 Thread Zhijiang(wangzhijiang999)
3. I suggest you upgrading the version to 1.5 or above. The checkpoint process may be faster than the current version. Best, Zhijiang -- 发件人：aitozi 发送时间：2018年9月18日(星期二) 14:53 收件人：user 主　题：Re: 回复：InpoolUsage & InpoolBuffers inconsist

回复：[DISCUSS] Dropping flink-storm?

2018-09-28 Thread Zhijiang(wangzhijiang999)
Very agree with to drop it. +1 -- 发件人：Jeff Carter 发送时间：2018年9月29日(星期六) 10:18 收件人：dev 抄　送：chesnay ; Till Rohrmann ; user 主　题：Re: [DISCUSS] Dropping flink-storm? +1 to drop it. On Fri, Sep 28, 2018, 7:25 PM Hequn Cheng wrote: > H

回复：Memory Allocate/Deallocate related Thread Deadlock encountered when running a large job > 10k tasks

2018-10-08 Thread Zhijiang(wangzhijiang999)
. Or I guess we can set the ExecutionMode#PIPELINED_FORCED to not generate blocking result partition to avoid this issue temporarily. Best, Zhijiang -- 发件人：Piotr Nowojski 发送时间：2018年10月4日(星期四) 21:54 收件人：Aljoscha Krettek 抄　送

回复：Small checkpoint data takes too much time

2018-10-09 Thread Zhijiang(wangzhijiang999)
backpressure. You can check the metric of "checkpointAlignmentTime" for confirmation. Best, Zhijiang -- 发件人：徐涛 发送时间：2018年10月10日(星期三) 13:13 收件人：user 主　题：Small checkpoint data takes too much time Hi I recently encounter a

回复：What are channels mapped to?

2018-10-11 Thread Zhijiang(wangzhijiang999)
will be consumed by the first reduce task. Best, Zhijiang -- 发件人：Chris Miller 发送时间：2018年10月11日(星期四) 16:54 收件人：user 主　题：What are channels mapped to? Hi, in the OutputEmitter, the output channel can be selected in different manner. eg

回复：Need help to understand memory consumption

2018-10-16 Thread Zhijiang(wangzhijiang999)
not recycled by gc, so you will see no changes in memory consumption. After you restart the TaskManager, the initial memory consumption is low because of lazy allocating via taskmanager.memory.preallocate=false. Best, Zhijiang -

回复：Need help to understand memory consumption

2018-10-16 Thread Zhijiang(wangzhijiang999)
The operators for stream jobs will not use memory management which is only for batch jobs as you said. I guess the initial feedback is for batch jobs from the description? -- 发件人：Paul Lam 发送时间：2018年10月17日(星期三) 14:35 收件人：Zhijiang

回复：TaskManager failure detection

2017-02-22 Thread Zhijiang(wangzhijiang999)
mponent in JobManager is in charge of recovering state from complete checkpoint, and the state would be set onto Execution in ExecutionGraph. Best, Zhijiang For yarn cluster mode,  --发件人：Dominik Safaric 发送时间：2017年2月22日(星期三) 19:05收件人

回复：How back pressure works in Flink?

2018-03-06 Thread Zhijiang(wangzhijiang999)
. BTW, from FLINK-1.5 release, the network flow control is changed to classic credit-based mechanism. That means the sender transfers buffers only based onreceiver's announced available buffers and will not send extra data any more, so there are no in-flight data accumual

回复：Checkpoints very slow with high backpressure

2018-04-07 Thread Zhijiang(wangzhijiang999)
improve barrier alignment, that has already been verified to decrease the alignment time greatly for backpressure scenarios. zhijiang --发件人：Piotr Nowojski 发送时间：2018年4月6日(星期五) 00:06收件人：Edward 抄　送：user 主　题：Re: Checkpoints very

回复：multiple consumer of intermediate data set

2017-03-13 Thread Zhijiang(wangzhijiang999)
case with JobVertex(A). Cheers, Zhijiang--发件人：윤형덕 发送时间：2017年3月13日(星期一) 12:43收件人：user 主　 题：multiple consumer of intermediate data set Hi All, figure1 https://ci.apache.org/projects/flink/flink-docs-release-1.2/fig

回复：multiple consumer of intermediate data set

2017-03-14 Thread Zhijiang(wangzhijiang999)
-B1,A1-IntermediateResultPartition-B2,A2-IntermediateResultPartition-B1,  A2-IntermediateResultPartition-B2 in the right graph. Cheers, Zhijiang-发件人：lining jing 发送时间：2017年3月15日(星期三) 10:54收件人：user ; Zhijiang(wangzhijiang999) 主　题：Re

回复：question about record

2017-03-27 Thread Zhijiang(wangzhijiang999)
buffers. Cheers,Zhijiang--发件人：lining jing 发送时间：2017年3月27日(星期一) 15:46收件人：user 主　题：question about record Hi All , data transmission is achieved through the buffer. If record serialised to multiple buffer, whether different records will

回复：PartitionNotFoundException on deploying streaming job

2017-04-04 Thread Zhijiang(wangzhijiang999)
when response the PartitionNotFound to track the reason. Wish your further findings! Cheers,Zhijiang --发件人：Kamil Dziublinski 发送时间：2017年4月4日(星期二) 17:20收件人：user 主　题：PartitionNotFoundException on deploying streaming job Hi guys, When I

回复：PartitionNotFoundException on deploying streaming job

2017-04-05 Thread Zhijiang(wangzhijiang999)
you find the reason for your case. Wish your sharing! cheers,Zhijiang--发件人：Kamil Dziublinski 发送时间：2017年4月5日(星期三) 16:07收件人：Zhijiang(wangzhijiang999) 抄　送：user 主　题：Re: PartitionNotFoundException on deploying streaming job Ok thanks I

回复：Changing timeout for cancel command

2017-04-12 Thread Zhijiang(wangzhijiang999)
Hi Jürgen,      You can set the timeout in the configuration by this key "akka.ask.timeout", and the current default value is 10 s. Hope it can help you. cheers,zhijiang --发件人：Jürgen Thomann 发送时间：2017年4月12日(星期三) 19:04收件人

回复：回复：Changing timeout for cancel command

2017-04-13 Thread Zhijiang(wangzhijiang999)
ack in hdfs.  cheers,zhijiang--发件人：Jürgen Thomann 发送时间：2017年4月13日(星期四) 15:32收件人：user 主　题：Re: 回复：Changing timeout for cancel command Hi zhijiang,I checked this value and I haven't configured it so I thi

回复：Yarn terminating TM for pmem limit cascades causing all jobs to fail

2017-04-19 Thread Zhijiang(wangzhijiang999)
 native memory, so you can try to upgrade the version as Stephan's suggestions.  Good luck! Cheers,zhijiang--发件人：Stephan Ewen 发送时间：2017年4月19日(星期三) 21:25收件人：Shannon Carey 抄　送：user@flink.apache.org 主　题：Re: Yarn terminating TM for

回复：Multiple consumers on a subpartition

2017-04-25 Thread Zhijiang(wangzhijiang999)
cheers,zhijiang--发件人：albertjonathan 发送时间：2017年4月26日(星期三) 02:37收件人：user 主　 题：Multiple consumers on a subpartition Hello, Is there a way Flink allow a (pipelined) subpartition to be consumed by multiple consumers? If not, would it mak

回复：Question regarding configuring number of network buffers

2017-06-07 Thread Zhijiang(wangzhijiang999)
it can help you. Cheers,Zhijiang  --发件人：Ray Ruvinskiy 发送时间：2017年6月7日(星期三) 23:59收件人：user@flink.apache.org 主　题：Question regarding configuring number of network buffers The documentation provides the formula #slots-per-TM^2 * #TMs * 4

回复：An addition to Netty's memory footprint

2017-06-30 Thread Zhijiang(wangzhijiang999)
the direct memory usage by netty PooledByteBuffer can be largely reduced and under controlled easily. cheers,zhijiang --发件人：Kurt Young 发送时间：2017年6月30日(星期五) 15:51收件人：dev ; user 主　题：An addition to Netty's memory footprint Hi, Ufuk

<    1   2