Return-Path: X-Original-To: archive-asf-public-internal@cust-asf2.ponee.io Delivered-To: archive-asf-public-internal@cust-asf2.ponee.io Received: from cust-asf.ponee.io (cust-asf.ponee.io [163.172.22.183]) by cust-asf2.ponee.io (Postfix) with ESMTP id C2AF1200BB6 for ; Fri, 4 Nov 2016 19:27:56 +0100 (CET) Received: by cust-asf.ponee.io (Postfix) id C11EB160B04; Fri, 4 Nov 2016 18:27:56 +0000 (UTC) Delivered-To: archive-asf-public@cust-asf.ponee.io Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by cust-asf.ponee.io (Postfix) with SMTP id 6E8F7160AEA for ; Fri, 4 Nov 2016 19:27:54 +0100 (CET) Received: (qmail 4729 invoked by uid 500); 4 Nov 2016 18:27:53 -0000 Mailing-List: contact user-help@beam.incubator.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: user@beam.incubator.apache.org Delivered-To: mailing list user@beam.incubator.apache.org Received: (qmail 4719 invoked by uid 99); 4 Nov 2016 18:27:53 -0000 Received: from pnap-us-west-generic-nat.apache.org (HELO spamd4-us-west.apache.org) (209.188.14.142) by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 04 Nov 2016 18:27:53 +0000 Received: from localhost (localhost [127.0.0.1]) by spamd4-us-west.apache.org (ASF Mail Server at spamd4-us-west.apache.org) with ESMTP id E133DC1444 for ; Fri, 4 Nov 2016 18:27:52 +0000 (UTC) X-Virus-Scanned: Debian amavisd-new at spamd4-us-west.apache.org X-Spam-Flag: NO X-Spam-Score: -1.3 X-Spam-Level: X-Spam-Status: No, score=-1.3 tagged_above=-999 required=6.31 tests=[DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, HTML_MESSAGE=2, RCVD_IN_DNSWL_LOW=-0.7, RCVD_IN_MSPIKE_H2=-0.001, RCVD_IN_SORBS_SPAM=0.5, RP_MATCHES_RCVD=-2.999, SPF_PASS=-0.001, WEIRD_PORT=0.001] autolearn=disabled Authentication-Results: spamd4-us-west.apache.org (amavisd-new); dkim=pass (2048-bit key) header.d=yahoo.com Received: from mx1-lw-eu.apache.org ([10.40.0.8]) by localhost (spamd4-us-west.apache.org [10.40.0.11]) (amavisd-new, port 10024) with ESMTP id IYDl2QN655Gy for ; Fri, 4 Nov 2016 18:27:48 +0000 (UTC) Received: from nm1-vm7.bullet.mail.gq1.yahoo.com (nm1-vm7.bullet.mail.gq1.yahoo.com [98.136.218.86]) by mx1-lw-eu.apache.org (ASF Mail Server at mx1-lw-eu.apache.org) with ESMTPS id 85A4F5F3F4 for ; Fri, 4 Nov 2016 18:27:41 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s2048; t=1478284054; bh=Ylqvo47QzRRTCbVee+34rjdI8Xq8L3rlq0GVurk8W28=; h=Date:From:Reply-To:To:In-Reply-To:References:Subject:From:Subject; b=eWKpD0CxHtItXi/YPdwGGg057w7jWZqZe+It0UBAC4uiHXa10JcCFyjGfDfPcBHEyzQf1UdqZuPOsRTcqRSDg+CQAEGR5qp+AlScBgBgIc9McSmAPB0W5IiLTrIhIC+zWLr763yOfGqN26lp7812dmGbtfvoRU7XyUlgqsdC1/HSZVx1VDBl7gQIeuSSs1N7xHMcMskwjDyk3gSTYX+3/8IbqhO1sIBB/8u5Pgl9Y5+5uSbEqiQd+42AHxHIZduecmmW3Fqfy6UTlPeBUZ1XOhdtbHXwVYY7q2PrwuKlcpS/oHIxGdiqP/qBHwjGOcMaa5rqpu915yy+ms1EYrn3gQ== Received: from [216.39.60.183] by nm1.bullet.mail.gq1.yahoo.com with NNFMP; 04 Nov 2016 18:27:34 -0000 Received: from [98.137.12.196] by tm19.bullet.mail.gq1.yahoo.com with NNFMP; 04 Nov 2016 18:27:34 -0000 Received: from [127.0.0.1] by omp1004.mail.gq1.yahoo.com with NNFMP; 04 Nov 2016 18:27:33 -0000 X-Yahoo-Newman-Property: ymail-3 X-Yahoo-Newman-Id: 994823.71672.bm@omp1004.mail.gq1.yahoo.com X-YMail-OSG: 2RHhtv4VM1mI4GQjvC.npbD1fmXgzh1p8N9Rbq_XA2_dJYOig3XY6GJJBvL0XvK qYTIHcOQ9S70.nU51_q7xDh6hlx0DEHLwKhv97dzEU0yzlvTNdtZ1dVsUGIF66.dNlhdu4F6apwD yF3p25T2Myoavdh5UCRTPGxvarQpRDlpA7zQjwsEy9.SCrqkG9Q.IeEYqhHb3tGJEbyCGMHjchdQ M0JWxAAXXOcZ8VYTyPg21GnSOdKHNONH9F0n7reVltRKwRVudG5vRXfJ1EcJUuQXGh7pFdBf6R72 MPEMgfu2pGMN0k_xnEROcstqmT1MEVRPsfgzNC8U1UAZPUeb.IScBtRQgdFHOmVY.vmIAOi2FiM8 nzXEHdUJbslHioGMbJ0R0ZvWHg6.4E5DCCNnWPhG57xXZta.RLdqJo.75_kHAmJqdQIbCxT67IDa ncEhKbhPkUPlUOU4dlPKRYFJb1CDnE7S6gPcJZ.ftFiY6h.T9h5nRlJHK2PP1YnmpnPSDSY1hASd eArI4FbWtzkQOw0tuviJFKMsimyxmxBPcGLbgqA-- Received: from jws300046.mail.gq1.yahoo.com by sendmailws124.mail.gq1.yahoo.com; Fri, 04 Nov 2016 18:27:33 +0000; 1478284053.594 Date: Fri, 4 Nov 2016 18:27:32 +0000 (UTC) From: amir bahmanyari Reply-To: amir bahmanyari To: "user@beam.incubator.apache.org" Message-ID: <15262555.513268.1478284053203@mail.yahoo.com> In-Reply-To: References: <1353304123.152534.1478145208604.ref@mail.yahoo.com> <1353304123.152534.1478145208604@mail.yahoo.com> <526822538.873219.1478202790312@mail.yahoo.com> Subject: Re: What does this exception mean to you? MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_513267_1414491071.1478284053203" archived-at: Fri, 04 Nov 2016 18:27:56 -0000 ------=_Part_513267_1414491071.1478284053203 Content-Type: multipart/alternative; boundary="----=_Part_513266_6963684.1478284053194" ------=_Part_513266_6963684.1478284053194 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Hi Ajoscha,Thanks for your reply.I am using Microsoft Azure VM A11 for a 4 = nodes Beam/Flink cluster & Dxy for Ingestion i.e. Kafka 2 nodes cluster giv= en below. After I create the partition on Kafka, the describe option show that its is= evenly distributed between the two Kafka nodes.Or, at least this is my und= erstanding. Below is partial output and it shows that all 2048 partitions a= re in use.Thanks so much for your help. I hope we can raw some conclusion o= ut of this and find the bottleneck.Have a great weekend.Amir- [aba@kafka01 kafka_2.11-0.10.0.1]$ ./bin/kafka-topics.sh --describe --zooke= eper kafka01:2181 --topic linroad3 |moreTopic:linroad3 =C2=A0PartitionCount= :2048 =C2=A0 =C2=A0 ReplicationFactor:2 =C2=A0 =C2=A0 Configs:=C2=A0 =C2=A0= =C2=A0 =C2=A0 Topic: linroad3 Partition: 0 =C2=A0 =C2=A0Leader: 1 =C2=A0 = =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topi= c: linroad3 Partition: 1 =C2=A0 =C2=A0Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replic= as: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partitio= n: 2 =C2=A0 =C2=A0Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: = 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 3 =C2=A0 =C2=A0Le= ader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2= =A0 =C2=A0 Topic: linroad3 Partition: 4 =C2=A0 =C2=A0Leader: 1 =C2=A0 =C2= =A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: = linroad3 Partition: 5 =C2=A0 =C2=A0Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas:= 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: = 6 =C2=A0 =C2=A0Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2= =C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 7 =C2=A0 =C2=A0Leade= r: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0= =C2=A0 Topic: linroad3 Partition: 8 =C2=A0 =C2=A0Leader: 1 =C2=A0 =C2=A0 = =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linr= oad3 Partition: 9 =C2=A0 =C2=A0Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1= =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 10 = =C2=A0 Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 = =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 11 =C2=A0 Leader: 2 =C2=A0 = =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topi= c: linroad3 Partition: 12 =C2=A0 Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1= ,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 13= =C2=A0 Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 = =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 14 =C2=A0 Leader: 1 =C2=A0 = =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topi= c: linroad3 Partition: 15 =C2=A0 Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2= ,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 16= =C2=A0 Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 = =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 17 =C2=A0 Leader: 2 =C2=A0 = =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topi= c: linroad3 Partition: 18 =C2=A0 Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1= ,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 19= =C2=A0 Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 = =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 20 =C2=A0 Leader: 1 =C2=A0 = =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topi= c: linroad3 Partition: 21 =C2=A0 Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2= ,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 22= =C2=A0 Leader: 1 =C2=A0 =C2=A0 =C2=A0 Replicas: 1,2 =C2=A0 Isr: 1,2.......= ...........................................................................= ...............=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad3 Partition: 2045 = Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: 2,1=C2=A0 =C2=A0 = =C2=A0 =C2=A0 Topic: linroad3 Partition: 2046 Leader: 1 =C2=A0 =C2=A0 =C2= =A0 Replicas: 1,2 =C2=A0 Isr: 1,2=C2=A0 =C2=A0 =C2=A0 =C2=A0 Topic: linroad= 3 Partition: 2047 Leader: 2 =C2=A0 =C2=A0 =C2=A0 Replicas: 2,1 =C2=A0 Isr: = 2,1 From: Aljoscha Krettek To: amir bahmanyari ; "user@beam.incubator.apache.org= " =20 Sent: Friday, November 4, 2016 10:44 AM Subject: Re: What does this exception mean to you? =20 Just out of curiosity, what machines are you running this on? I'm asking be= cause the number of task slots should roughly correlate with the number of = CPU cores. On Fri, 4 Nov 2016 at 10:34 Aljoscha Krettek wrote: You should try and find out why everything is just happening on one node. H= ave you looked at your Kafka Partitions, i.e. is the data evenly distribute= d across the partitions of your Kafka topic or is all data pushed to one pa= rtition? This would actually explain why processing is only happening one o= ne node, namely the node that is reading the partition that has all the dat= a. On Thu, 3 Nov 2016 at 20:53 amir bahmanyari wrote: Thanks Aljoscha.I have been tuning Flink memory, NW buffers etc.And this oc= curred in THAT ONE NODE that I see *.out logs get created by Flink.I lowere= d the memory that Flink allocates i.e. 70% by default to 50%.=C2=A0And this= exception was thrown in that one node only. Other nodes were up & didnt cr= ash.There is SOMETHING different about THAT ONE NODE :-) I cannot figure it= out.At every ./start-cluster, THAT ONE NODE may/may not change on random b= asis.So I cannt just tune THAT ONE NODE. Next time, another node may become= THAT ONE NODE. I have the followings set in flink-conf.yaml in each node: akka.ask.timeout : 300s jobmanager.heap.mb: 256 //Could this be too small?=C2=A0taskmanager.heap.mb= : 102400taskmanager.memory.fraction: 0.6 //Changing this to a lower value c= auses the exception below. Am testing with 0.6 <0.7 default. taskmanager.numberOfTaskSlots: 512 taskmanager.memory.preallocate: false parallelism.default: 2048 taskmanager.network.numberOfBuffers: 131072 Appreciate any feedback.Amir- From: Aljoscha Krettek To: amir bahmanyari ; "user@beam.incubator.apache.org= " =20 Sent: Thursday, November 3, 2016 12:45 AM Subject: Re: What does this exception mean to you? =20 That looks like a Flink problem. The TaskManager on beam4 seems to have cra= shed for some reason. You might be able to find that reason by looking at t= he logs on that machine. On Thu, 3 Nov 2016 at 04:53 amir bahmanyari wrote: Thanks+regards,beam4 and beam1 are hostnames.BenchBeamRunners.java is my Be= am app running in a four servers FlinkCluster. Other nodes are still running except the one that failed beam4.beam1 has th= e JM running. Amir- java.lang.RuntimeException: Pipeline execution failed=C2=A0 =C2=A0 =C2=A0 = =C2=A0 at org.apache.beam.runners.flink.FlinkRunner.run(FlinkRunner.java:11= 3)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.beam.runners.flink.FlinkRunner.= run(FlinkRunner.java:48)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.beam.sdk.= Pipeline.run(Pipeline.java:183)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at benchmark.fli= nkspark.flink.BenchBeamRunners.main(BenchBeamRunners.java:622)=C2=A0 =C2=A0= =C2=A0 =C2=A0 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Metho= d)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at sun.reflect.NativeMethodAccessorImpl.invok= e(NativeMethodAccessorImpl.java:62)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at sun.refle= ct.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43= )=C2=A0 =C2=A0 =C2=A0 =C2=A0 at java.lang.reflect.Method.invoke(Method.java= :498)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.client.program.Package= dProgram.callMainMethod(PackagedProgram.java:505)=C2=A0 =C2=A0 =C2=A0 =C2= =A0 at org.apache.flink.client.program.PackagedProgram.invokeInteractiveMod= eForExecution(PackagedProgram.java:403)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.a= pache.flink.client.program.Client.runBlocking(Client.java:248)=C2=A0 =C2=A0= =C2=A0 =C2=A0 at org.apache.flink.client.CliFrontend.executeProgramBlockin= g(CliFrontend.java:866)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.clie= nt.CliFrontend.run(CliFrontend.java:333)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.= apache.flink.client.CliFrontend.parseParameters(CliFrontend.java:1189)=C2= =A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.client.CliFrontend.main(CliFro= ntend.java:1239)Caused by: org.apache.flink.client.program.ProgramInvocatio= nException: The program execution failed: Job execution failed.=C2=A0 =C2= =A0 =C2=A0 =C2=A0 at org.apache.flink.client.program.Client.runBlocking(Cli= ent.java:381)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.client.program= .Client.runBlocking(Client.java:355)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apac= he.flink.streaming.api.environment.StreamContextEnvironment.execute(StreamC= ontextEnvironment.java:65)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.beam.ru= nners.flink.FlinkPipelineExecutionEnvironment.executePipeline(FlinkPipeline= ExecutionEnvironment.java:118)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.bea= m.runners.flink.FlinkRunner.run(FlinkRunner.java:110)=C2=A0 =C2=A0 =C2=A0 = =C2=A0 ... 14 moreCaused by: org.apache.flink.runtime.client.JobExecutionEx= ception: Job execution failed.=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.fli= nk.runtime.jobmanager.JobManager$$anonfun$handleMessage$1$$anonfun$applyOrE= lse$7.apply$mcV$sp(JobManager.scala:714)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.= apache.flink.runtime.jobmanager.JobManager$$anonfun$handleMessage$1$$anonfu= n$applyOrElse$7.apply(JobManager.scala:660)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at o= rg.apache.flink.runtime.jobmanager.JobManager$$anonfun$handleMessage$1$$ano= nfun$applyOrElse$7.apply(JobManager.scala:660)=C2=A0 =C2=A0 =C2=A0 =C2=A0 a= t scala.concurrent.impl.Future$PromiseCompletingRunnable.liftedTree1$1(Futu= re.scala:24)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.concurrent.impl.Future$Pro= miseCompletingRunnable.run(Future.scala:24)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at a= kka.dispatch.TaskInvocation.run(AbstractDispatcher.scala:41)=C2=A0 =C2=A0 = =C2=A0 =C2=A0 at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTas= k.exec(AbstractDispatcher.scala:401)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.co= ncurrent.forkjoin.ForkJoinTask.doExec(ForkJoinTask.java:260)=C2=A0 =C2=A0 = =C2=A0 =C2=A0 at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.pollAndEx= ecAll(ForkJoinPool.java:1253)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.concurren= t.forkjoin.ForkJoinPool$WorkQueue.runTask(ForkJoinPool.java:1346)=C2=A0 =C2= =A0 =C2=A0 =C2=A0 at scala.concurrent.forkjoin.ForkJoinPool.runWorker(ForkJ= oinPool.java:1979)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.concurrent.forkjoin.= ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:107)Caused by: java.lang= .Exception: The slot in which the task was executed has been released. Prob= ably loss of TaskManager 06dff71ba6ab965ec323c8ee6bf3d7d1 @ beam4 - 512 slo= ts - URL: akka.tcp://flink@10.0.0.7:44399/user/taskmanager=C2=A0 =C2=A0 =C2= =A0 =C2=A0 at org.apache.flink.runtime.instance.SimpleSlot.releaseSlot(Simp= leSlot.java:151)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.runtime.ins= tance.SlotSharingGroupAssignment.releaseSharedSlot(SlotSharingGroupAssignme= nt.java:547)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.runtime.instanc= e.SharedSlot.releaseSlot(SharedSlot.java:119)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at= org.apache.flink.runtime.instance.Instance.markDead(Instance.java:156)=C2= =A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.runtime.instance.InstanceManag= er.unregisterTaskManager(InstanceManager.java:215)=C2=A0 =C2=A0 =C2=A0 =C2= =A0 at org.apache.flink.runtime.jobmanager.JobManager$$anonfun$handleMessag= e$1.applyOrElse(JobManager.scala:847)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.r= untime.AbstractPartialFunction.apply(AbstractPartialFunction.scala:36)=C2= =A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.runtime.LeaderSessionMessageFi= lter$$anonfun$receive$1.applyOrElse(LeaderSessionMessageFilter.scala:44)=C2= =A0 =C2=A0 =C2=A0 =C2=A0 at scala.runtime.AbstractPartialFunction.apply(Abs= tractPartialFunction.scala:36)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.fli= nk.runtime.LogMessages$$anon$1.apply(LogMessages.scala:33)=C2=A0 =C2=A0 =C2= =A0 =C2=A0 at org.apache.flink.runtime.LogMessages$$anon$1.apply(LogMessage= s.scala:28)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.PartialFunction$class.apply= OrElse(PartialFunction.scala:123)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.= flink.runtime.LogMessages$$anon$1.applyOrElse(LogMessages.scala:28)=C2=A0 = =C2=A0 =C2=A0 =C2=A0 at akka.actor.Actor$class.aroundReceive(Actor.scala:46= 5)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at org.apache.flink.runtime.jobmanager.JobMan= ager.aroundReceive(JobManager.scala:106)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at akka= .actor.ActorCell.receiveMessage(ActorCell.scala:516)=C2=A0 =C2=A0 =C2=A0 = =C2=A0 at akka.actor.dungeon.DeathWatch$class.receivedTerminated(DeathWatch= .scala:46)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at akka.actor.ActorCell.receivedTermi= nated(ActorCell.scala:369)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at akka.actor.ActorCe= ll.autoReceiveMessage(ActorCell.scala:501)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at ak= ka.actor.ActorCell.invoke(ActorCell.scala:486)=C2=A0 =C2=A0 =C2=A0 =C2=A0 a= t akka.dispatch.Mailbox.processMailbox(Mailbox.scala:254)=C2=A0 =C2=A0 =C2= =A0 =C2=A0 at akka.dispatch.Mailbox.run(Mailbox.scala:221)=C2=A0 =C2=A0 =C2= =A0 =C2=A0 at akka.dispatch.Mailbox.exec(Mailbox.scala:231)=C2=A0 =C2=A0 = =C2=A0 =C2=A0 at scala.concurrent.forkjoin.ForkJoinTask.doExec(ForkJoinTask= .java:260)=C2=A0 =C2=A0 =C2=A0 =C2=A0 at scala.concurrent.forkjoin.ForkJoin= Pool$WorkQueue.runTask(ForkJoinPool.java:1339)=C2=A0 =C2=A0 =C2=A0 =C2=A0 .= .. 2 more =20 =20 ------=_Part_513266_6963684.1478284053194 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Hi Ajoscha,
Thanks for your reply.
I am using Microsoft Azure VM A= 11 for a 4 nodes Beam/Flink cluster & Dxy for Ingestion i.e. Kafka 2 no= des cluster given below.
3D"Inline
3D"Inline

After= I create the partition on Kafka, the describe option show that its is even= ly distributed between the two Kafka nodes.
Or, at least this is my understanding. Below is partia= l output and it shows that all 2048 partitions are in use.
Thanks so much f= or your help. I hope we can raw some conclusion out of this and find the bo= ttleneck.
Have a great weekend.
Amir-

[aba@kafka01 kafka_2.11-0.10.0.1]$ ./bin/kafka-topics.sh --d= escribe --zookeeper kafka01:2181 --topic linroad3 |more
Topic:linroad3  Parti= tionCount:2048     ReplicationFactor:2     Configs:
  &nb= sp;     Topic: linroad3 Partition: 0    Leade= r: 1       Replicas: 1,2   Isr: 1,2
  &= nbsp;     Topic: linroad3 Partition: 1    Leader: 2 =       Replicas: 2,1   Isr: 2,1
        Topic: linroad= 3 Partition: 2    Leader: 1       Replicas: 1,2 &n= bsp; Isr: 1,2
        Topic: linroad3 Partition: 3   &nbs= p;Leader: 2       Replicas: 2,1   Isr: 2,1
     = ;   Topic: linroad3 Partition: 4    Leader: 1     =   Replicas: 1,2   Isr: 1,2
        Topic: linroad3 P= artition: 5    Leader: 2       Replicas: 2,1  = ; Isr: 2,1
        Topic: linroad3 Partition: 6    L= eader: 1       Replicas: 1,2   Isr: 1,2
     =   Topic: linroad3 Partition: 7    Leader: 2     &= nbsp; Replicas: 2,1   Isr: 2,1
        Topic: linroad3 Pa= rtition: 8    Leader: 1       Replicas: 1,2  = Isr: 1,2
        Topic: linroad3 Partition: 9    Le= ader: 2       Replicas: 2,1   Isr: 2,1
      &n= bsp; Topic: linroad3 Partition: 10   Leader: 1       Re= plicas: 1,2   Isr: 1,2
        Topic: linroad3 Partition:= 11   Leader: 2       Replicas: 2,1   Isr: 2,1
  &nb= sp;     Topic: linroad3 Partition: 12   Leader: 1   &nb= sp;   Replicas: 1,2   Isr: 1,2
        Topic: linroa= d3 Partition: 13   Leader: 2       Replicas: 2,1  = Isr: 2,1
        Topic: linroad3 Partition: 14   Leader:= 1       Replicas: 1,2   Isr: 1,2
        = Topic: linroad3 Partition: 15   Leader: 2       Replica= s: 2,1   Isr: 2,1
        Topic: linroad3 Partition: 16 &= nbsp; Leader: 1       Replicas: 1,2   Isr: 1,2
    &= nbsp;   Topic: linroad3 Partition: 17   Leader: 2     &= nbsp; Replicas: 2,1   Isr: 2,1
        Topic: linroad3 Pa= rtition: 18   Leader: 1       Replicas: 1,2   Isr:= 1,2
&n= bsp;       Topic: linroad3 Partition: 19   Leader: 2 &n= bsp;     Replicas: 2,1   Isr: 2,1
        Top= ic: linroad3 Partition: 20   Leader: 1       Replicas: = 1,2   Isr: 1,2
        Topic: linroad3 Partition: 21 &nbs= p; Leader: 2       Replicas: 2,1   Isr: 2,1
    &nbs= p;   Topic: linroad3 Partition: 22   Leader: 1     &nbs= p; Replicas: 1,2   Isr: 1,2
.................................................= ................................................
        Topic: linroad3= Partition: 2045 Leader: 2       Replicas: 2,1   Isr: 2= ,1
&nbs= p;       Topic: linroad3 Partition: 2046 Leader: 1   &n= bsp;   Replicas: 1,2   Isr: 1,2
        Topic: linro= ad3 Partition: 2047 Le= ader: 2       Replicas: 2,1   Isr: 2,1



From: = Aljoscha Krettek <aljoscha@apache.org>
To: amir bahmanyari <amirtousa@yahoo.com>; "us= er@beam.incubator.apache.org" <user@beam.incubator.apache.org>
<= b>Sent: Friday, November 4, 2= 016 10:44 AM
Subject: = Re: What does this exception mean to you?

Just out of curio= sity, what machines are you running this on? I'm asking because the number = of task slots should roughly correlate with the number of CPU cores.
<= br clear=3D"none">
On Fri, 4 Nov 2016 at 10:34 Aljoscha Krettek <aljoscha@apache.org> wrote:
You should try and find out why everything = is just happening on one node. Have you looked at your Kafka Partitions, i.= e. is the data evenly distributed across the partitions of your Kafka topic= or is all data pushed to one partition? This would actually explain why pr= ocessing is only happening one one node, namely the node that is reading th= e partition that has all the data.

On = Thu, 3 Nov 2016 at 20:53 amir bahmanyari <amirtousa@yahoo.com> wrote:
Thanks Aljoscha.<= /div>
I have been tu= ning Flink memory, NW buffers etc.
And this occurred in THAT ONE NODE = that I see *.out logs get created by Flink.
I lowered the memory that = Flink allocates i.e. 70% by default to 50%. 
And this exception w= as thrown in that one node only. Other nodes were up & didnt crash.
At every ./start-cluster, THAT ONE NODE may/may not change on rand= om basis.
So I cannt just tune THAT ONE NODE. Next time, another node= may become THAT ONE NODE.

I have the f= ollowings set in flink-conf.yaml in each node:

akka.ask.timeout : 300s
jobmanager.heap.mb: 256 //Could this be too small? 
taskm= anager.heap.mb: 102400
taskmanager.memory.fraction: 0.6 //Changing t= his to a lower value causes the exception below. Am testing with 0.6 <0.= 7 default.
taskmanager.numberOfTaskSlots: 512
taskmanager.memory.preallocate: fa= lse
parallelism.= default: 2048
ta= skmanager.network.numberOfBuffers: 131072


=
Appreciate any = feedback.
=
Amir-
<= /div>
From:<= /b> Aljoscha Krettek <aljoscha@apache.org>
To: amir bahmanyari <amirtousa@yahoo.com>; "user@beam.incubator.apache.org" <user@beam.incubator.apache.org>
Sen= t: Thursday, November 3, 2016 12:45 AM
Subject: Re: What does this exception mean to you?

That looks like a Flink problem.= The TaskManager on beam4 seems to have crashed for some reason. You might = be able to find that reason by looking at the logs on that machine.
On Thu, 3 Nov 2016 at 04:53 amir bahmanyari &l= t;amirtousa@yahoo.com> wrote:<= br clear=3D"none" class=3D"yiv0897504002gmail_msg">
Than= ks+regards,
beam4 and beam1 = are hostnames.
BenchBeamRunner= s.java is my Beam app running in a four servers FlinkCluster.
Other nodes are still running except the one that failed beam4= .
beam1 has the JM running.

Amir-

java.lang.RuntimeException: Pipeline = execution failed
      &nbs= p; at org.apache.beam.runners.flink.FlinkRunner.run(FlinkRunner.java:113)
        at org.apache.b= eam.runners.flink.FlinkRunner.run(FlinkRunner.java:48)
        at org.apache.beam.sdk.Pipeline.ru= n(Pipeline.java:183)
      =   at benchmark.flinkspark.flink.BenchBeamRunners.main(BenchBeamRunners= .java:622)
        at = sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
        at sun.reflect.NativeMethodAc= cessorImpl.invoke(NativeMethodAccessorImpl.java:62)
        at sun.reflect.DelegatingMethodAccess= orImpl.invoke(DelegatingMethodAccessorImpl.java:43)
        at java.lang.reflect.Method.invoke(Me= thod.java:498)
       = at org.apache.flink.client.program.PackagedProgram.callMainMethod(Packaged= Program.java:505)
      &nb= sp; at org.apache.flink.client.program.PackagedProgram.invokeInteractiveMod= eForExecution(PackagedProgram.java:403)
&n= bsp;       at org.apache.flink.client.program.Client.runBloc= king(Client.java:248)
     =   at org.apache.flink.client.CliFrontend.executeProgramBlocking(CliFr= ontend.java:866)
      &nbs= p; at org.apache.flink.client.CliFrontend.run(CliFrontend.java:333)
        at org.apache.flink.c= lient.CliFrontend.parseParameters(CliFrontend.java:1189)
        at org.apache.flink.client.CliFr= ontend.main(CliFrontend.java:1239)
Caused = by: org.apache.flink.client.program.ProgramInvocationException: The program= execution failed: Job execution failed.
&= nbsp;       at org.apache.flink.client.program.Client.runBlo= cking(Client.java:381)
     = ;   at org.apache.flink.client.program.Client.runBlocking(Client.java:= 355)
        at org.ap= ache.flink.streaming.api.environment.StreamContextEnvironment.execute(Strea= mContextEnvironment.java:65)
   =     at org.apache.beam.runners.flink.FlinkPipelineExecutionEnvir= onment.executePipeline(FlinkPipelineExecutionEnvironment.java:118)
        at org.apache.beam.run= ners.flink.FlinkRunner.run(FlinkRunner.java:110)
        ... 14 more
Caused by: org.apache.flink.runtime.client.JobExecutionException: Job = execution failed.
      &nb= sp; at org.apache.flink.runtime.jobmanager.JobManager$$anonfun$handleMessag= e$1$$anonfun$applyOrElse$7.apply$mcV$sp(JobManager.scala:714)
        at org.apache.flink.runtime= .jobmanager.JobManager$$anonfun$handleMessage$1$$anonfun$applyOrElse$7.appl= y(JobManager.scala:660)
    &n= bsp;   at org.apache.flink.runtime.jobmanager.JobManager$$anonfun$hand= leMessage$1$$anonfun$applyOrElse$7.apply(JobManager.scala:660)
        at scala.concurrent.impl.F= uture$PromiseCompletingRunnable.liftedTree1$1(Future.scala:24)
        at scala.concurrent.impl.F= uture$PromiseCompletingRunnable.run(Future.scala:24)
        at akka.dispatch.TaskInvocation.run(= AbstractDispatcher.scala:41)
   =     at akka.dispatch.ForkJoinExecutorConfigurator$AkkaForkJoinTa= sk.exec(AbstractDispatcher.scala:401)
&nbs= p;       at scala.concurrent.forkjoin.ForkJoinTask.doExec(Fo= rkJoinTask.java:260)
      =   at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.pollAndExecAll(F= orkJoinPool.java:1253)
     = ;   at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.runTask(ForkJo= inPool.java:1346)
      &nb= sp; at scala.concurrent.forkjoin.ForkJoinPool.runWorker(ForkJoinPool.java:1= 979)
        at scala.= concurrent.forkjoin.ForkJoinWorkerThread.run(ForkJoinWorkerThread.java:107)=
Caused by: java.lang.Exception: The slot = in which the task was executed has been released. Probably loss of TaskMana= ger 06dff71ba6ab965ec323c8ee6bf3d7d1 @ beam4 - 512 slots - URL: akka.tcp://= flink@10.0.0.7:44399/user= /taskmanager
      &nbs= p; at org.apache.flink.runtime.instance.SimpleSlot.releaseSlot(SimpleSlot.j= ava:151)
        at or= g.apache.flink.runtime.instance.SlotSharingGroupAssignment.releaseSharedSlo= t(SlotSharingGroupAssignment.java:547)
&nb= sp;       at org.apache.flink.runtime.instance.SharedSlot.re= leaseSlot(SharedSlot.java:119)
  &nbs= p;     at org.apache.flink.runtime.instance.Instance.markDead(Ins= tance.java:156)
       = ; at org.apache.flink.runtime.instance.InstanceManager.unregisterTaskManage= r(InstanceManager.java:215)
    =     at org.apache.flink.runtime.jobmanager.JobManager$$anonfun$ha= ndleMessage$1.applyOrElse(JobManager.scala:847)
        at scala.runtime.AbstractPartialFunction.= apply(AbstractPartialFunction.scala:36)
&n= bsp;       at org.apache.flink.runtime.LeaderSessionMessageF= ilter$$anonfun$receive$1.applyOrElse(LeaderSessionMessageFilter.scala:44)
        at scala.runtim= e.AbstractPartialFunction.apply(AbstractPartialFunction.scala:36)
        at org.apache.flink.run= time.LogMessages$$anon$1.apply(LogMessages.scala:33)
        at org.apache.flink.runtime.LogMessa= ges$$anon$1.apply(LogMessages.scala:28)
&n= bsp;       at scala.PartialFunction$class.applyOrElse(Partia= lFunction.scala:123)
      =   at org.apache.flink.runtime.LogMessages$$anon$1.applyOrElse(LogMessa= ges.scala:28)
        = at akka.actor.Actor$class.aroundReceive(Actor.scala:465)
        at org.apache.flink.runtime.jobm= anager.JobManager.aroundReceive(JobManager.scala:106)
        at akka.actor.ActorCell.receiveMess= age(ActorCell.scala:516)
    &n= bsp;   at akka.actor.dungeon.DeathWatch$class.receivedTerminated(Death= Watch.scala:46)
       = ; at akka.actor.ActorCell.receivedTerminated(ActorCell.scala:369)
        at akka.actor.ActorCell= .autoReceiveMessage(ActorCell.scala:501)
&= nbsp;       at akka.actor.ActorCell.invoke(ActorCell.scala:4= 86)
        at akka.di= spatch.Mailbox.processMailbox(Mailbox.scala:254)
        at akka.dispatch.Mailbox.run(Mailbox.sca= la:221)
        at akk= a.dispatch.Mailbox.exec(Mailbox.scala:231)
        at scala.concurrent.forkjoin.ForkJoinTask.doEx= ec(ForkJoinTask.java:260)
    &n= bsp;   at scala.concurrent.forkjoin.ForkJoinPool$WorkQueue.runTask(For= kJoinPool.java:1339)
      =   ... 2 more





------=_Part_513266_6963684.1478284053194-- ------=_Part_513267_1414491071.1478284053203 Content-Type: image/png Content-Transfer-Encoding: base64 Content-Disposition: inline; filename=blob.jpg Content-ID: <8f8e631a-798b-2e36-8e39-fe0e1f0f09de@yahoo.com> iVBORw0KGgoAAAANSUhEUgAAAJEAAAC5CAYAAAAoJ914AAAgAElEQVR4Xu2dB1QUVxfHf/QioiIB BFEULKjYY4kmaizYosbYa4y9d2PDLnY0KtbY0NijsXejsYD1UzQWbAiCoqiwIG0X+M7M0gVZ2MWA zOTkHGHf3Lnv/37cd+fNvLtanJbFIx2SAmoooCVBpIZ60qmiAhJEEghqK5AEUbO9d9Q2JhnInwpI EOXPcddoryWINCpn/jQmQZQ/x12jvZYg0qic+dOYBFH+HHeN9lqCSKNy5k9jEkT5c9w12msJIo3K mT+NqQ2RReeqbG5ojE7oO2ZMfIBnGh0NbK2Z1Lc4tWKCaOn6POFTbRwal2VSqyJE/ePJ0L8yEb+s Ha4/mWKa2CxcxroVvnh/jjETr62H17xHHNTI9UwZMNyOIne8WXBOIwb/cyNqQmTAiOk1aKAfDWY6 PNxylcleCX0yK8qIn+1oYG+AiTbgH4Cz63Msqpdmwk/mVDDTRQd4fOLypyFqWIaVdeFkikGs3MEO u72+ykEVPneS5xxUEkSZQqoeRA5l2TrWjDf7nhHTwp4yz5/w029ByotWKceWHsYEnIvArHVRSiVA VHdAdUYXi+JgoD49qxtnClHlDo4MMH3PsI2v0u+MBFGmg5zTDdSCqFLfGiypEsPvI+7AuDr0K/FB /PeeVF6XxH21DQ4JECV91K4KJ5wLZAqRGGnSRKJEGyJgZfSSTMoe+TJ5r4y0v/f3TJw6EqeS11DX AlvxTDnef95nnU+iGSt+nZT4Gfg/klGoTIrpLO3UGvQ6CfA2v1SmjkxGaBlTbIlMip7C75tZJtgP kuFdwFSazpRyfIXr4jJUeu1Hx4UviG7qxMH2BXh64CqjjselwEhNiCAZihQDlnSBjyKRKQN+McNr ozJnEoEqFpkw3SkhqmySZoALJOZYSoBIA13K9pU7lKGO9yMldCJQRvgmQKiEJdm24KP4O5JBSwQ8 GeycjhM5bz/7kahOBfb0LsTrE1cY+lcc6BVn2ZISlH0VwI+uz4lO8l19iBJNJf5FpxqAzKazVDlN Okltys/Ts/XJnCi1PWUkUkZD5SFAWYT3qSKdlFgnodFRmL7shYw5zREXwUGXW7i/S/y95iASLab5 6083sRanQKMUjiVGh09D5Jte/vURRKmnO+EiiVB/BFG6AEoQKQcmMeqEhnLQOyJ5sKyL0Ka8IYH/ 3KDfjsRYpGGISDMIaaNH2p/VjUSpcrK0kSU7kSjtlJnz001OXyFb05lBcyf2tDUh6NxN+u9KnrjQ K4n7Mhsc3r1huMsjlLmqehClykFUiERp7+aUOYgiIcnNZDpLSJpDM8qJ0kaWNO0/ns4ScrAPUk6U BmQDhk6rThvLyDTTltBMm18m16azbQznltxg3mP1IVLenaWcmtLeTSUmy6C8OzNOdXclC4oESxIW CzOBKAnSxIVNOd6ekdjVTb47S3WnFR6JP0aQsHD4MUSCwWT/hJ9kj17jW8xCujvL6fAo2c9bCmRr OstbXZS8zWkFJIhyWuF8YF+CKB8Mck53UYIopxXOB/YliPLBIOd0FyWIclrhfGA/CSKfMlq5trtF CxnkWt8kx1LsxZcgknDIrgJSJMquctJ5SQpIEEkwqK2ABJHaEkoGJIgkBtRWQIJIbQklAxJEEgNq KyBBpLaEkgEJIokBtRVQD6KoF9y/8CceHieIbb+chT86pHIo+vlptq7bzMVHrwmTg55JTfrMdcW5 RNb8llass6bX526dTYhusmv0Ag76vicqwWP7XqtSQRTiuZSJS4/xRseGqrWrYGkovBuqoGSXcRqG SE7QjV2sdt/FJV8Zcj1TKnScwvx+dZL37n9uVfPZ9bIJ0XnW9dxGUJPONNXdy6K9T0kFUdh5Vg2Z yxm9ZkxYMo7aRdRTNeNIJMdn92hGrnlE8bZDGd6mJvov/mTFriKMWdkDe/UuK52togLZhCjZ+vP9 Qxjj8TgVRK+OjGXo7760nLqTvjWStzmr6NNHzTKESHaCWR0XEdRnBcu6OJJ4pXfet5BVropddi8o nZclBXIAolDOL+jIcq/vGTTDDu+127n8Mgo9k3LU7zWI/k0rktVn8hlBJDsxhTYLDJl+0oVGuun0 W+HL3+7zWHr0ATK5HqaOLRj96wgaldDjyc7+9F1nRqu2Lzh7wInJZydS8dY2lrtt5dyLKIyKf8eA iZP5sYIR7664M3PRQW6/iwKjNsw/MpY6WZL5y26cAxDdZseg8ewN0kXP0IzytWtiTQA3L9/mjVyX OiO2Mr5R0SypmhFEIgjnGrBhTXpTVyiX3PowxbMWLkuGUK/IS84uncTCJ61ZveUX9EWIXuE8eT3j mlih92If4weeotoiN7pVgPtbRjD4bF1Wb7HnmPM0Xg7ehms726Rol6UOfOGNcwCi86z6cS5nKEjD sesYXl8JjOLGCvrNOURYuX4sm98poSKHaupmBFHQgTF0XlWMuSfGU+8jU+dxEwZ/yE4WtS2m/NR7 Ld1GXaftuvXUvC5AZM/ssxP5FvDdPZif19xLY6UFs892Qza5P4tuF6d5z5781KoeZQqqP0Wr1vO8 0SoHILrK5p5TORTejAn7x1E7SYdEuNL+PnOhMsyJ7m2k17D9OLpsYlIj8zSGlBCFjDnILOdCKSC6 RbsNq6lxJR2I/nBk8YER1EzrkkLGI8+DHN27j/2PHRi/diGtimfud35pkQMQveLEzF6su/UNIzxm 0KBgopQJEJm0Y9rWIVTJgsIZ350lTFkndKnXbQg9m5XF8PVdzl+LoUG/b/Ff1AOXq/VSTWdLX3dk zcquIE5nyZEIEcjtFOy7mPmdq6IXdJk/Lsnp26lBkqfyJ38wuv8mSs86zZj6WejAF940ByAC5d3Z Hey7r8C1QzmEnDfk4nxGLTlLVL3xrB/XlCS2VBD4k4uNilfc3OvO8h2e+AormkZFqNJiNJOHNcAy 6g775y9mk6cvMrkhtjU7MezXvtQuSkJinQIi5Pj/vZx5y45xL0yOkZkTDQaOYGLTN7i3nc6eRNtt JjB94DeYqeB3fmmSIxChuMrmAVM59N6QUnW/p6x+QmJNeXq5LqGtQ9ZyCmnFOnfjmDMQCX1+f5MD 61ez//pzwuSGFK7QmD4DBlC/ZMriDKqJI0Gkmk7/VSu1IfocjksQfQ6Vs38NCaLsayedmaCABJGE gtoKSBCpLaFkQIJIYkBtBSSI1JZQMiBBJDGgtgISRGpLKBlIgii+cVYeREjCSQokKyBFIokGtRWQ IFJbQsmABJHEgNoKSBCpLaFkQIJIYkBtBSSI1JZQMiBBJDGgtgISRGpLKBmQIJIYUFsBCSK1JZQM ZBmi6Je3uXTodzxOx/LjglW0LZUoopzwp54c27uJvzwrMiJpz5mc4NuH2Ll7F573lFVEDAtXpFHf Ufxcv6S4EySzQ3o9NjOF/tvPVYfozmbGue3jWUhiMRkHerkpIXpzZi4umy/xJlyR0JsUGxSf7WTC mI0EFK5ItdolMZU94fL1h4TJzXCeuJYBtRM2Fn5Ch09DJJSW+YtNHn9w/s57ItHD1KI8XV1W0rUi XFjUAJdjCcb1TLGr24fJU9tTVhV6/9uxyTNXVx0iz6X8vDOIxj+1QG//XPb4JkP0fP8YZl23osNP ZXkwexUXSQnRIfY8rkibpqWTCjmE/D2DQcsvI68zls2/Ome6B+1TEL37Zz4DZvxNkYTSMhaGITw+ v5dnxafRs34CRD7dcZvVGuugU8z+dSPGKbdW55mhyr2Oqg5RUh8ec2DcEDyeJEOU3D0Vt0onRKcn Ku7Lz2ppmZRyi5HoSf+Eog8+7BrUn7MN17O2S9ncOyp5zLP/BKLESGT4wxw2/1IrU8myXVqGhEiU AJHNvS2MHX2K6os20bdy1jZQZupkPm7w2SFSPN7JtMkbeUh5fl6wnB+SEvOMR0HV0jLKmkPKL1Kn xQzOjW+UOicC7FpMZfrwppQSyv9Jh0YU+IwQCXdpHrjN3cVDLHAe/RsD6qpWp0jV0jLysCDehvtx dNo4PMqmgCgxJ4ry5ciSaey0GM+f05ylmo4aQShbX1WVnZwokkf7pzDD4y5RhpXoMnkKHZ1UA0jo Z4Y50eNtDBywC7tUpWWUec9q+xQQJeVE6RVy0JCS+djMZ4hEcp7uH8tkjwdQrjNTxvbC6aus5SOf LC3zW39mHDWixYhRtK9uiWHobTbNns+Jqp+IRKbD2ebaFst8PPCa7HrOQ/RyH9OGrOFfm87MWdYX x2ysz3x6nSiYW/vWsmn7P8qainqmWJavQ/vuPelcq0SanMgQ2+96MnZ4D6qqHgg1qfcXaSvHIfI/ MIJRmx+ATU2cK1mkEdGehoN+ILObbWnFOnezl+MQJZYoTl8G1UrvSRB9cRB9/g5JEH1+zbNyxWxE oqyY10xbCSLN6JhTViSIckrZfGRXgigfDXZOdVWCKKeUzUd2JYjy0WDnVFcliHJK2XxkV4IoHw12 TnVVKi2TU8rmI7sSRPlosHOqqxJEOaVsPrIrQZSPBjunuipBlFPK5iO7EkT5aLBzqqt5GiK57Bl3 Ll9g/47tHLMfz/VpjZU6yZ9xYskc5h64T0iMHjXHbWBjV4ec0jDf29UIRIHhMYREK3gXpSBKEUdo TCyF9HUw1tWhqKEuBfS1KV4wq99BnfnY+GwZzNirFtj4n+Jyzdl4ixCFcG5uT8Y9bsHyqZ0pVziK MHkB7KwKZ25QapEtBbIFkTwunr/9Qrn7NoJ7byOIkMd9dHE9bS0cChuhpaX8SF9HC6sC+tia6FO/ uCnC55o5HrK5Rx/cyiZA5LeLvj+ep8n+VXQtoZkrSFY+rUCWIRLg2ePzVow8nzqsTfQpbKBLfDyE y2ORxcQSFhNLXHy8+Pt2DmY0s9NEdEgNUeCekTQ/YMsQh2tsPuQPJRoyZu5MOlXI2uYACRzVFcgy RN2OJGwOTHMN/9cxPPSPQhahIFoej4GeFibGOjhYG2JfLPVOwXjixbPH1bSmuqWJ6t6m2zI1RD5b +tBheTDt5yxneK1ojk8byIKXPfhjX3+c1LySdHr6CmQTImEqUoKQeMg+xPIkMIawSDkhH2IpaKSD iZEu9sX0MTNNf4uHnYEBrk1Kqjk26UB0/Dt27eiDo2D5zFwqT4ClN6aQkHareT3p9LQKZBmibzbc wc5KmSQLeU1BfR3xf2M9HQS0IhVxPAtNLD+TseD+b2JobF2IGc1s1RyV1BCFHP6V79aVZMvBIVQT btROzqDGJD0JIjVV/tTpWYZIa6wnRQrq0qBCQSpaG2Goq5PK/nNZNB/kseleM0YRx8u3cu77RfE+ TMGmLvb8/HXabURZ7W2axDr0GBOaz+dlfzcWOMOxqWNYbTmR0/NboIkMLKve5Yf2WYZo8J/P+N0r CEVsnBh5rMz0sSqsRwFDHQwNtfgQE4uxgS6R0bEo4iAiKpYYeTxvQhUiOMKhq61FvzqWrPjRTvy3 ekcaiIDg6x7Mn7uFk35QsnFvpk3oxdfm6l1FOjtjBbIMkWDqcXAUS86/5PiD9/i+i1ZZXzszA5qX L8LYBsVwMJfKcqgsXC5vmC2IImL88X27jeCwy/iHRHL/tR1BH8yIkhsQFK7cn6ylpUPxQnqYGMRg UeADTlZvKF44DmP94pQw64yxvrq5UC5XNh+5l2WI7r2cx4v3+4mPzyDviZUTpYiiaIFyFDQoRmxc NDGx74hRvCc2LjIZsCI/4mg1QYRNOvK2AlmG6MS/NdLtcWB4EP5hL4mLT1691tHSxsbECpuCVume U85qAnZFO+dBBSN5GxCBsU1RjPKg95p2OfsQpV4mIh59wuQxRCoiCZO/Rk9LFwNtfYoaFUZPJ/3V Yl39mjQuszb7ffLZw4RZR/ETLehjal2Wxh270/5rKzS/Pv2IXRNdedV8EyNt/2TS7Kd8u3A8LdPe XJ5zp8txC2bN75hpoYrsdzx3nZlliNZdtsLG1Ap9bQN0dQqiq22Cnk5BdLSVf5PRird8iPb9ZC8V 8bEEyF5R1mowDUtPz74iIkSvabltKA2JxN9zL6vWXqfU4MUMqJ0+Rpc2DueYyQjmdCqTxeumgKjh J06VIMpc1xmntNDR0qFUYUcsTewxNfgK7cSnrMQTGnmP2LiPFxuFaS5c/oH3kTKCIoKJjY+lW9VD lP2qdeYXzahFKoiUjW5tGcuKmM5s6K+Dx7hteMkVfAgzpHavX3GWbWTBX/eJ0C1M4TqdWNk8lMXz TvCUaD5E29B69AQ6OqaAT+7L8dXu7LgaDDYVKBV1jyI/bmIkidGmHbEHl7P6mC9x+saUdB7CeOPD CZGoEYErZrOd9rh21efA0u1ckemjb1SV3gt78nX2e53rzsxyJLoZ8Dtnn7oQHvlK7Iw2uhQwNMFY zwQd4T+tmKRORsfGIESdGIWcCEVkUr5kYmDFd3ZTqFVimHqCpAORz+6pTAtqys7htYiMNMLICIKP LGTYTUfcXH7AZ0UfjlpOZqEQieSRRGKEkZ6cG5smskohwJdczTbk1FKGnbFmyuzOOIb+jdtkD3S6 pYSoKv+buJKQtvMZWDchOxIjkTn9q9/Dw7sSE6d3xPGSO10u2LJoYhtsNT/PqqehBs7OMkTCNRVx Udx88TtP35/B9905ohQhmbpiqFsYO7OGOBR1pqr1z+hqa2Cd6COI5FxeM5JdBYfwWycdjq7azcmA aAh7y6tCTcQ8JTAlRMGeeKw+xO0wiAx5wzunHuwc3iCpLwKQq+RdWNa9EpBiOkuKRB0x89yO++6L +BlXpVv/PjT2XUeX3Y8xi4ij2tDF9P9aD+T3ObpqB3/+G0nZJl3o26kGX9LaZ7YgSktM8IcHhEe/ Qhb9gpjYcMJjgjDRt0RfxwRTg+IY65tjYSIMhIaPlBDJZTw4v521uyNp4TqaSpddGfO4NpvGNOaD EIkulEqC6NBXE1jSxZEr64eww3gIy7qXEyPRovA2qSAKPOjKmMv2TM8wEiUmz3Lu/DGdue+c2el0 V5zOxjd/zdrd0G3WUBomEhN5hkX9/6bEtDl0zqw8nIalyklzGoEoJx38pO2Ud2e6RhSzr0Xrnt1p bKcHPn8xY/Ex/I0sqOVQkL8DlBBZXFjH5A2eyL7uy8rKD5m26SqKIvZUsA3hoo5zKoiECLJn3goO +MgxrVQHh3cX0WmZcjqrxcOFyzkREENMjA1NR4ylY9C6hJyoHbG75zBfmNJaRLJ2921iYmIwqNKB KYMaSJHoP4NGunCuVCBvR6JcKWn+c0qCKP+NucZ7LEGkcUnzn0EJovw35hrvcRJEQU7qvhymcd8k g3lEASkS5ZGBys1uShDl5tHJI75JEOWRgcrNbkoQ5ebRySO+SRDlkYHKzW5qBKLw8HCePfMlVCYj KlL5HrVwaOtoU8yqGHZ2JTES3snQ8KGx0jIPPejabQ3/JvnXKmmzo/zVeVb+Oo8/7srQTbWvX87L c2sYP3Mv3jJdSjYexMK5HUn5OpKGu5trzakFUVxcHLe8vQkMCEzqYIECBcR/xyoURMfEEB8fL1YG sbEuTnnHchgaaK7EjMZKy9xcRav+L+l/cBDKt4kMKSK+P+3H7qE9WWs5ic2jqnBv2QDG32qj3Nfv t4dBnT2wmubOmGoPcOs9g//9tIYDAyrn2sHOKceyDVF0dDRXr10jNFRGIVNTrG2ssbKyooCxcZKv coWCgIAAXr0K4tGjRxgYGNCqVctUbdTvmAZKy2S0X1+MUKdpstWDvhUAEbb/0Xn3eupc7EOHFHv+ /+fegd7X27F/Uw/s1e9UnrKQLYjCwsO5cuUqUVFR2Ba3wcnJSYw2IZHPMTYwJyTyKe8inlBA35Li prXQ1tbDx8cHrytXKVSoEC1btkBPNxvf45mutCqWlilzncWtx3KhzUr2Dq3OhzNzaTNVxvhDC2h+ U9ivf1K0rm/qSPtJUxnfrBR6aeGSn2RKnRnoLbzMt5e+YTSJhbXy957/bEHkdeUKwcFvcXQsj33p 0ryPfMqNF+uIVLynbsnRBMqu4fX8N3FQjPSK0q3qQXS0Dbh7919u3LxJGQcHqlevhqGhBt5uRPXS Mrbn5vPTpJcM3Nab1+NH8/SXP3FrbY5cFsQbcYt3CFdWTGT6KVsmCUWyHqWtKHKGGTVcIB2I8nP1 kSxDFBAQyG1vbywtLKhRozqx8XJO+UxAHhchQlOz+EDCol5y7qlyF0elYt2wNqnOi1Av6peayJWr 13jw4AG1an2NY/nyGgjbWSktE8wJlx7MvapHjEMvdrt35KNiaqGHGfO9K6YLLzODNBBJkSjd8coy RFevXef9u3dUrlwZcwtTXoXfprBhSW4ErKewoR0VLDugr1NAfO9aFhWAnrYx66/WIkIeTK/qp7At 1JBdu3ZjbGxM+x/baRyiT5eWUdZznHg8AkXt0Rxf3PrjNwxFiBZhtuQ8U02FHOgaHbdv4udyqXMi p+Md6O3Zhr3beon7y6ScSChZ1bigSgN6+MhRLC0t+bpmDfxDPXkddoeChjaUNW+V7vn3Xu/l4L/9 iFKEUsmqMx2cdvLPhYs8e/aM5s7NRFvqHaqXllGcmUvnqTJGbf2Wv3uuxGjmNuY2M+fJucME2dal XOFX/LNoItOv1GHpX1NoXMibNe0HsafqNDYPrcBd96G4BPVinxDBvNfTts9Bqs1xZ3ile6zoN49X /baypmP+KxSZ5UgkQCTkQUI+dOfVdmIU4RjpmYkRKPGIfeODzlfJb6L7vDksRiJtLV0qF+uBz6NH eHp6Uad2bcqVU/eNdRVLy3CGGd1n8aTnOrb2KIfPtv5022DD3D0zcLo0mS7zz4nligvb12PwVBe6 Vlaua8nv7WHSlDWc9FNQuFIHXJcMp7744r2c+3umM2H5OZ4rTKncZSrLRtb/ot6dVvWPO1sQVXB0 pHTpUmIyLRxFC5TFrohyW2jk/sEYPV1DbKw28fHaH/kRp2XA+x9OcPzaUzGnqlSxoqq+Su1yqQLZ gqh6tWpYWxdLWEgUNuUnwxLuuZM3q3qiX7Iq2oYfF/XUMjQhqvVczl+9zbf161G6dOlcKo3klqoK ZAsiBwd7ypcrx0vZTaJjZeJGRJtCtdESa6dBXFR4ugAlOvXgwUOuXL1K0yaNsba2VtVXqV0uVSDL EJ0+fUa8s/rmm7qERPoSEqUs3mBVsCrCLldVDmGd6eFDH35q/yMmJuqWIFblilKbnFQgyxAJjzpe v35Do0YNMTTSIzD0GnHxCjES6Wpn/lwsKOg1Fy5ewNz8K777tj7a2h/nTSp3OHHzYrn24j77xJgm 7qHf4s3X/TYx8lMVPDK80Hl+63EKq8SdqsE32LVpJ0dvBxMtbJKs6Mwvw9riZATi3v+DAaIlXWNb GvUZQd+65gTf2MOGbaf535sY0LWjk8t02n+hz0OyDJGfnx/ed+5iZmZGndq1iImViXvzhSINqhxC FPLxeSQCZGdnp8opGbcRIfqbt8bFaDXVhfbi3bUf+6avwys6ABuhlpC6EJW6z3YXN+449mVIpzrY 8pgj7ivYEdkEV5cfiEoqINGAyBtbmLhWQbd1Nbk/ci3vfpjG8CY5UStJPdk0fXaWIZLL5dy5e4/A wABKliyBUyXV99gLt/ZeXlewtLTAuVkz9fuSsBffqcVzrtOL34TCC3e3M3JPLNV0zxL6rQDRjY9K zAypdAe3yccwGzKXn82OM2PGQ+osHEfzpCoLKSLR+w303a7P8N96UjXR49CTzBt6CYeZM6lyI7EK iVAIQjjvLNYz+6LjMZtrpRPA0/xbMOprp0ELWYZIuHZsbCyXLl9GJgsT14yEtZ5PTUviKyO3bnHn 7r9iPtWsaRPxQazaR2JBh2kWHF0cQjv3/hTcPYE/C3bA8fJqZVWzhpHplpjRP7GYsVdL0rbwRR6W dWGSc8o6HSkgCnSny60KbB3VKEX1NaFCyBb4ZQ7VbiVC9A2vTq9mylFzRrl1w0mYAn/fxiEfY77p 0Z/+39vlQPU2tRXUiIFsQSRcOTIyksteV4iMiBCTY7uSJbEpbpPq6fyHiAhe+L/gydOnvHnzmq++ sqBRwwaae0EtqSrID7ybPp+gpt1g+zXs57Xm/byE0nj1hLIuH5eYKctdto1dwmGaMWVJ1zTf+5EC oqA19NhbgHEfRaLrVFg4GceLiTmRLgWsytOm7wjapngzLfLxUVa4naBw74yrt2lkJP9DI9mGSPBZ eF/o3r37vH8fwvPnvujo6FC0aFG0tbWQyxW8ffs2qWvCsoCwQi200diRorRM1VNLGbnjAQa1fsZ9 kDn7EuordpSlX2LGTIhEt2xxjrnMq3oujPk+g0hU6jabf13JwyrKqclK/pCT61axI7YVy35txrsU OVG6/ZI/Ze+c2Tz6ZimTnFW7e9WYPp/JkFoQJfoovB7r5+fPcz8/IiKUT/OFQ5i6hKf99valNTN9 pRUlVX0iT1YP3Y/JiIX0rJSiIJV1OiVmhhiyfdoVHCfPoX3EVsatek8b1xE0Ti8nKgvyVxfxWLWL f56Gi3dndtVa031Ay+S7M7EyW3JxLPDk98GbOR0m3JkJd3MtGDDqhy/21VmNQPSZgJcuk0sVkCDK pQOTl9ySIMpLo5VLfZUgyqUDk5fckiDKS6OVS32VSsvk0oHJS25JkSgvjVYu9VWCKJcOTF5yS4Io L41WLvVVgiiXDkxeckuCKC+NVi71VaMQxcXLufBsPj7BR7A3a0r9Ur+K3+8hHV+2AhqDKDz6JZef L+XhmwMY65mL+8zKmDendonRFDFS8w3GL3sM8nzvNAJRmACQnxs+rw/iYO5M0zKLOPVoPI+DT+Bg 3pw6JUZSxEjaGpTnacmgA2pDFBYdgKfvUh4GH8LRoj21bIdhamgjfnXVVX93vF9uw67Id3xnPw0z oy/0TfUvlQ4V+6UWRML3mwlT2KM3h9HW0qGMeSuKFihDDZsB3AhYx9sPj7gduJkP8mBx+7Twfa9m xg4quiY1yysKZBsiWZQ/nn5L8XlzJFVfbQQyebIAAAeISURBVArVon2lrRx5MJT7Qfv4EBNEPMqv rq5crDsNSk+nqHFWv8Q3IzmV9YIejtjEjt5C2Q7NHmdmpS5klWRdLH71mDGJ1UJUuuzHNQNUOk2F Rn5n1nEktin9hMJcKrTXdJNsQRQa5YeXCNDRj/xJhOjQ/YHcDtyKIi65EKjQ2MmqGw3tBZDULeQg WJMgEgpzHZ/6EzMiRrLPrU3S3jtNg/Ipe1mGKDTqOV7Pl+ITfIxipjV4KbuRZN+8gCON7Gegq1OR 2LgHHL4/BP+Qi0mfC5VD7gXtxcmqKw3sp2NurG70kCD6nLBkdK0sQ3TgXl/83l8U78LqlhjL1pvK /WNCOb1OlY+x5T48C42igJ4O42sW5HcvB6JilV80PKr+M7z8fsPLbxklCtenV41Tan6hcEYQyfE9 uZSp847gLZOnrsPIbf4YupC1N58pS8mkKBcj9z3Joulu7BPKDZfrSjuLHWw3S67LmHY66zu+OtfW KksQl+81k1ViaZmM7KepLXnOjSmuR7jxNhKMbWk2YjrzOlbg2ZY+dFhekpGj37LF/QYhumXoM38J o+sJL4AHc22LG66bL/FEKHvcdgJrp9lxskcf3Moq/RSnYJ9ujKl0gTV7/FFY1WPCIlc6VdCD4Gus mTmTzZffib/vUj0Ij2dNkgp1ZRfILEO04lI5ShdtQotyv4n1hlIeXoFhbH/wBll0LDFxcXQuZ05b h6If+Xb4/mCuv1hDmwrrqW7TL7u+ZzidhQi1Gcde4ts5yxnT0IzAEwsZOvsJHTftZFBlP+7eiKVE 5VKYKs4zp+UkHvb1YGuPWDb36MNGq7Gsm9YS68CjzBi8hNONMoLoCKU6zWHliLpwcwMTRuzBdvYB FraUfdJ+4mCH3L/NC4uyVCqqy9NtI2nnXoLlnhOxFiEKprnLIqZ+r8uhSX1ZENFfrEqrt2ck7d0i +HnxZHo5hbFj2S0aTavN5bQQHavIgMUu9LL3ZfWQieyp4cr1KVU5M6sLo31as2ZeX6pxk3WTxrOB Qf8NRMKaT3snD/S0C/D47XERAgMdU0IUdVh96xUfFLHi7/pVtsLG8AKR8mDxZ2E6E76tevP1Rgjf YJ1TEJ2b24ARgcM47t4xIUe4yco2w7j0k5CAF+Xazo1sOHAV31dvCJTJqSgk5vWv0bPTNqqsPM64 ukquVU+slVHmRNNP2O+N2CYRIiHq/b5qH4cfvOBtwDsiUBZgtxUhckgqxu4j/Ly5Mmv/7oz/0A6s LL2Ef8YqHZQ/fIhfObjwUSRKBkPsQ9Bojrt/xca6k3kxZi9rOiqrFoi2TzX6/BAdvNef5+//ERcP m5SZxx7vzkmRpF3FTXgGleN/r8OoYmFCY5sQVns5igUfRKhqefLXv31EgHJyOhMhej+efxa3RrnT SwnRlU4eLDBeTZsVBZnu8Ss/2PolD36dK3TtdpB66/cyrLo4RJxwacB4vYyns+S7s3ts6NqP0803 saTguvTtp4KoIEudR3H1J3fW9qmG6T/JBUbThUiEqjMBfXqxq+ZKjgwVHUw4Uk+TyuksDUTizzZs r+GCbOZR3ForVbm/sRedz/4H01nKxDrx8YbgkCIuWqzL2NhhtrjgeNV/JWcfuSTlQ0IbE30rwmNe KRPr0tMxL6CZxPpe32Usa2ujFNXQDONbbrSZ4MX3KaezBa/pvXk9dbz60OHPKqzfNRTHB1sY2G8j iEsEoeLAHq42AffJjfjq0T5GDVvP7RYZQXSSKr+44tqzEh/OCtOlN63XHKLVvYzsp4xEiEsTbyf9 iVsrOcddBzDl6LcZR6KEyGSxrgvdNxVkgDid6XD3n6dYti6WeSQSIWrC3ZFdcZV1Zvms7lSK9cRt 4Cz2ffUfTGfCOKV3iy+sBwkQCWX3fq75Nztv/8iD13+l+ItR/jMnbvH3pbxKW2HQv+HOztlMWysk oHKMS9Si97ipDBaS0+AzuPZ3ZacflOzYmyZ31+AlTkPlCL6+ninjt+GZ8F0d7QyW8luGkeghP3Sz 5MbeSwTqJifGehnaTwnRd9zfNoaB7jeIMKzBiJ5FWeyulylEjeUvOb9qJjP3ehMcYfSJxDq9SNSL Ur6HmT16KX8l1J8cXOMm867+B9NZ4ngJi41e/st5+PogMbFhCM/PhEOoY93acQ2H7w/i+ou1qSDS /GLjR4xKv8iCAp5LmjPwbg+1v0oiy3dnKX0UHnt4Pl+G1/NlxBOLka454xsGoq2lh/BayLKLpZFF vRBPkR57ZGF0c6ipz74VXLZsTfvqZry7+Qcu47ahGKT+ar9aEAl9FR7Anns6kxsv1mNqWJwx3/on SbDysqOYRFcp1lNcXJQewOYQHSqaDTw+l6ELTopTvLA29U2PIUz/pQHF1HxWojZESpACE0Bax3el p1LZqru4Mn32iQtVrHuJD16lV0FUHOk82EwjEAn9Fl5KEyJSyjyoqnVvEaDCRqXyoDSSy6oqoDGI hAsK60FC9LkVuJlKVl343n629HqsqiORh9tpFKI8rIPkuhoKSBCpIZ50qlIBCSKJBLUVkCBSW0LJ gASRxIDaCkilZdSWUDIgRSKJAbUV+D/K/QuQE/A+xAAAAABJRU5ErkJggg== ------=_Part_513267_1414491071.1478284053203 Content-Type: image/png Content-Transfer-Encoding: base64 Content-Disposition: inline; filename=blob.jpg Content-ID: <2dfaccc7-9cbd-6f76-022f-f229c0fa407e@yahoo.com> iVBORw0KGgoAAAANSUhEUgAAAIkAAACyCAYAAAB2meynAAAgAElEQVR4Xu2dB1QUVxeAP8rSFERU RBHFLvYuxhg1dmPU2HuJwRqxd8GOHSsW/O29RKPG3iv2gkasEUFQBFFAabvAf2YWcEHAXRYVZOac nMjOK/fd++19982+d0enyYCLcUiXpIE0NKAjQSLx8TkNSJB8TkPSfSRIJAg+qwEJks+qSCogQSIx 8FkNSJB8VkVSAQkSiYHPakCC5LMqkgqkGxLLTpVZX98EvZBgpox7gEcyXRraFGR830LUjA6ghctz 5V1ZLvoNL0nLIgYYEkvI6zcsc3nMOfnnDGHF2PGW2CQWi+DYrMfs/1y1DLkv9J2bt3954f4oIxo0 o98QW3Lf9WTOmYxo78u3kU5IDHGcXI16BlFgocfDDVeZcDleWIs8OPa2pV5xQ3LqAr5+NBUhMaTn uMp0s4nh3jFv/jEtxNA6JvD0OZ3m+xGV2lhL2eLSzhhvVSPVt6Wfv7fSaPVLsqyCHPel3nh+EX1J kKQPkhKl2DTSgsA9z4huXpySz5/SbnGA0kSVSrOhuwl+Z8KxaJmHogmQFCnJ+nH5yOH1mA5LAkVo BjtXpVX+D2wZ7MnG1Az8OQg+d19rcCRI0gVJ+b7VWFApmv853oVR9vxR+IP4711JDFIEtxXWlEiA 5NeKHGqRg6eHPBhyQFmw9sDqTKkIHiuvM+VOKtZMyZPEF63Y3o5+JWWJFUMfezNhd6jSu9Q2/vRz oNXvFbEPfY13AUsq5lQW8fVQdf3K6SDhXujjUEJKGqtMN8mmvvehiV5MlMcsAs8cZlTMKccz3vsJ fTbJHy9OQKh4/zufbvLhMr8k5V/70GHuC6IaV2B/2xz8t+8qw47Eqlg6KSSl+lRlaU3dpEC0qcTR prK0IRFaTDC6ikESO0rBk7T6vSSsjY9ZxLokxjBKg300oBI0Rfx9JSC2L+Nhi4dKtbwgy1gex8cT SmCIh0zZFolwCDKK/fGaP9e+EkVOADspmFq7uy/agOaexL4su3rl4vXRKwz+OxZkhVi0oDClXvnx m8tzldgiKSQpeg11IUnmORI9RgJAacYkSacLpSf5CAGo3EeIf2RcThIUpz3dqLan9CRvE4FI0nZi 0JsNAtcOwvRSXIhIk12x4ex3uo1bcMLnyaYbEQhj/t1zmRHHlWWq9auOSxW4uOQ607zU/TIk/fam GLiKU5QZZolNJnX9qUJSMKUgODkkSacjoYsEaD+BRJQjOXTfOyQJXiMkhP2e4R+tWjA3rcoY4X/u Bn9sS1inJIOkUhm2DbAAzwd0WSGQpMvvE2rRyep9CvFM2sAk8QbJp5tPDKOlJ0kSE31qYM09STLI 1f1ufMNyGk03hs0qsKt1TgLO3MRhh8qiVVYEt0XWlAgOZIjTY5SeNRkkmDFmejkaWsi5d8yXU3kL 4FDdmA+3HtLNPdH9fKqKJDGAcPszniQ5NGJMop8kiEzVkzxStp0rIQD+JCZJ7lWSlv90uon3Oh+y TUySsGSNSDatCIaL9wo20ZxZcINZT1KCBMhfgCkDbKhhpY++QoHPfV+mrnjJi7S+JZ9MHamvRpRu nySrEwIi8M2vn7g6STMmEelOunrx9XgNtT8+TEu6oorAN8CYXPExzqeQCO0lXy0pV1bf+ermG/o9 qetvogGNpptvIqHU6TfXgATJNzdB5hdAgiTz2+ibSyhB8s1NkPkFkCDJ/Db65hJKkHxzE2R+AXQe PQ/LtCf48uQyzPwazAYSSpBkAyNrO0QJEm01mA3qS5BkAyNrO0QJEm01mA3qS5BkAyNrO0QJEm01 mA3qS5BkAyNrO0QJEm01mA3qS5BkAyNrO8QMgOQN/x5cw+btZ3j0XoEsZxFqtO1Pr1+rk1dfO/Gk J67a6S+jamsJiRyvLf2ZtPsFssK1aFu/HK8vbOX0f5FYNHHCdWBdTLWQNG1I5ATc2MEKtx1c9A5F LjOjbIeJzP7DXmWXvBadS1UTNaAlJOdx7zido/KaDF47g59zA4rLrOnjzKH3Vem3ajZNLdOv7dQh kfNo53CGrnxModaDGdKqOgYv/mLpjtyMWNad4unvUqqZggYyBhKLTsxY2Rc7sYMn7Bs1iI1PvyAk oUeZ1mEeAX2WsqizHQkHPYM9bxNasTK2kqkzVANaQiLnztoeTDsgx37APByb2hB2YTHjFxyDhhOZ 92c9zLUQNzVPEnp0Iq3mGDH5mBMNUop7FN6cdpvFwkMPCJXLMLNrzvCxjjQoLOPpdgf6ulvwS+sX nNpXgQmnxlHu9maWuG7izItIjAv9RL9xE/itrDHBV9yYOm8/d4IjwbgVsw+OxF6L8WTVqlpCIgw7 gBvuE3A57JuoA+MK/Znu3I6iXyhwFQ19ph5rVqY0tYRw0bUPEz1q4rRgEHVyv+TUwvHMfdqSFRt+ x0CE5BVNJ6xmVCMrZC/2MLr/carMc6VrWfDa4MjAU7VZsaE4h5s683LgZlza2CR6q6xqaG3k1hKS N9xcO465B/wwr9Ca1o1tiL51gJ2nn2JSZzTTRzXGSgvpUvMkAftG0Gl5AWYeHU2dT9o/i6tg3EHb mde6gPKu5yq6DrtOa/fVVL8uQFKc6afGURfw3jmQ3ivvJ2ulOdNPdSV0ggPz7hSiWY8etPulDiVN P2Yw0GJYWa6qdpA83IDjuC34VR/Kmom/xE8tcv7d2BfnvUHUGbaNEfVypVspqQau99fS88+92Dmt Y3yDvMnaV0LybsR+pjWN71uE5DZt1qyg2pUUINlix/x9jlRPLqkilMce+zm0ew97n5Rg9Kq5/FIo 3cPJshW1g8RjPu3mHqN4z+XM/a3ERyWk9rmGakp9dRM/pRzVp07XQfRoUgqj1/c4ey2aen/UxXde d5yu1kky3Sx83YGVy7qAON189CSIwG3FtO98ZneqjCzgElsuyunbsV6itPKnWxjusI5i004w4kcN B/EdFNcOklvL6T3tb8KSeZKb7p2ZeTiM0r1X4tK6WLrVlOZzEsUrbu52Y8k2D7zD5GCcm0rNhzPh z3rkj7zL3tnzWefhTajcCJvqHflzbF9q5SE+cFWBBDm+p5cwa9Fh7ofJMbaoQL3+joxrHIhb68ns Smi71Rgm9/8Bi3SPJutW1A4SnnBgnCPrH4JpqVr8UDQXUf7XuXj3NfLczRm3eDg1tHiaJj1xzRxg aQkJEPkfF3a4s/uEJ77vFWCUm7K1O9Gxy69UyKddoCdB8r1A8gXHIUHyBZWrQdPaexINOtO0qASJ phr7MuUlSL6MXr+rViVIvitzfpnBSJB8Gb1+V61KkHxX5vwyg5Eg+TJ6/a5alSD5rsz5ZQajExcX l2mzCnyZIUutaqoByZNoqrFsWF6CJBsaXdMhS5BoqrFsWF6CJBsaXdMhS5BoqrFsWF6CJBsaXdMh S5BoqrFsWF6CJBsaXdMhS5BoqrFsWF6CJBsaXdMhawbJs+2MGbGWp8l7Kf47rvM7U0T1c2Hv68al bD7/kEBh76vMlKp95jGxufq756WdaZqa88uUTx8k1tVpWl4lXYBFNX7rWJd8CTK+PYv72DkcDdSn YOUfqJDfCAghpkgPBmYYJELqib9Zt3ELZ+++JQIZZpZl6OK0jC7l4Py8ejgdjhdIZoZt7T5MmNSW UloePf0yZsjcrWoGyZMtjBq9gYj283DrVimVkYVwZdnvzD0po+EYNwbVzpNuDaTlSYLPzabflNPk jk89YWn0jidnd/OskDM9foyH5FE3XKe1pGDAcaaPXYuJ6tHPdEuV/SpqBok6J/Ne7sF50Ep8mk/j f/3s0eaLmyokqaSeUDWf6EmeOsQfKn/EjgEOnKq/mlWdS2U/K2s54gyHJOz0FHovuUT9AS4UvuPG Tg8/ImWmlPqxD7/3a0lJYeZR80p36gniPUk8JNb3NzBy+HGqzltH34ranQVSU/TvqphGkAQeHscA 95sfFSAcxGrggEPPRhSON77Xlp5M2v0KmcwI8zI/ULUg+N88x91ABTL7kawc21TtnCWpQZI89YQy 50j8K7ybT+HM6AZJYxLAtvkkJg9pTFENIP2uLK3FYDSD5Mom9t56o+wu9DlXbv3Lu0iQlf6DObM7 iqubK8uaMPckmNYbx6JhPyuBSEyRVYbei5fwa2H1JFY39YQ8LIA373045DyKjaVUIEmISSK9ObjA me2Wo/nLuamUU0099SeW0giST9p+e4W1Tk4c9LOi/YyN4qriztp2TDsQRsMxxxhU+2ONBHiSf56W vKnGJE8207/fDmyTpJ5Qxh0riqtAkhiTpHRQXENNZePi2kGi4jkSjJ8wJdk77mJ0g4+5SZSQmPKr 81/0rqKextNMPbHYgSmHjGnuOIy2VfNjFHKHddNnc7RyGp7EbAibXVqTX73upVLxGtAIEoVCjr6+ SuD39izLR87k5FuVJHrxq5t/i/di3uxuFBOWN29PsdRxNmei6jJsjRN11cw0kPbDtCBu71nFuq3n lDnNZGbkL2NP22496FSzcLKYxAibn3owckh3Kqd/RZ5todEIkisr2rHtdQ3KCg/HIl9x98pN/CNJ lrM1IdleMEbF6lCvpGF84Aqley5k2m+l1V4WS09cMweXGkHyaM9wXPfGP2ZHH9PCFWnYrh8dfypG 0rflveHfvW6s3nNZTEdhZF6OBn2G0OOTcmkrQYIkC0LytUWWIPnaGk+5P408ydcWWYLka2tcgiRz aDwLSiF5kixotK8tsgTJ19Z4FuxPgiQLGu1riyxB8rU1ngX7kyDJgkb72iJLqSe+tsazYH8SJFnQ aF9bZAmSr63xLNifBEkWNNrXFlmC5GtrPAv2J0GSBY32tUXOkpDIQ59x99J59m7byuHio7nu3FCp N/kzji6Ywcx9XryLllF91BrWdlF5WdPX1u530p9WkPi/j+ZdlILgSAWRilhC5DHkkulhItMlt5E+ OfT1KGyWdKdJRujt0YaBjLxqibXvcS5Vn46nCMk7zszswagnzVkyqROlzSMJk+fA1kqb94lmhLRZ vw2NIJHHxnHaJ4R7b8K5/yaccHnsJxrQ19WhhLkRujo64j2ZHljlMMQmp4y6hXIh01V+rv31kPXd ++BaKh4Snx30/e0sjfYup4uau/G1lyF7tKA2JAIcux69ET1HWlfBnAaYG+ojZId9L48hNDqGsOgY YuPixM/blLCgiW1GfLuTQuK/ayjN9tkwqMQ11h/whcL1GTFzKh3LSoextEVZbUi6How//JSsxxdB 0Tzxi+R9eCxhETGYmehjaqJLUStDClsmnWriUOYVHlW9IFXz59RS9qSQPNrQh/ZLgmg7YwlDakZx xLk/c152Z8seBypo2VN2r64xJDoknS4CQ6J56h9NSLiCyOhYDGW65DDSpZS1EZa5U/4W2xoa4tIo SaKKdNghBUiO/MSObX2wE1o7OZOKY2DhjYnEh7Xp6EOqImhAbUh+WHMXWyulZxDiDVMDPcwM9Mgp 00MIPyIUsTwLifysVn0Do2lYMBdTmth8tmzaBZJC8u6fsfzkXoQN+wchHOuRH5tCtfEyCRIttawR JDojPchtqk9dO1MqWBtjpK+XpPvnoVF8kMekKFK0IpaXb+R4+UTyNkzBinZFGfCDNu8eF7pJFriG HGZMs9m8dHBlTlM4PGkEK/KP48Ts5mqfPc4AfX6XTajtSXpsfcLmG0EQF4uwQLHOY0i+XPrkNNLD 0FCH99ExmBjqExEVgyIWwiNjiJbHERiiEMFIuNpXysO27iURVkHaXckgAYKub2T2zA0c84EiDXvh PKYnNZK/gFy7TrNlbbUhEbRz2+8DM0/6ceJRCO8i0l7lqGrT3FifZmXMGdugIJWtc2RLRWflQWsE SXi0L95vNhMUdol7r3R5EGjL2wgzgj7kJkJhwNtwM/Lk+ICZoQmWpuHkMoqgcoHX2FlGYGJYiMIW nTEx0DYWycrqzpqyqw3J/ZezePF2L3FxKccdilgFMbGxmBnZksOwADGxUUTHBCNXhKCI/SBqR0dH j0K5f8POaoz4b+nKGhpQG5Kj/1ZLcUQB4UF4h7wgNu7j01c9HV2sc1phbZpycFraagy2eTplDQ0l kTKCN37hmFjnwTgLSp9ekdMBSdIXbcXE6hIc9YGomCgiFaHo6epioCMjr7EFhvoGKcqlb1CdhiVX pVdmeLSLMdMO4SO2YIBZwVI07NCNtjWsyPjnq4/ZMc6FV83WMdTmL8ZP/4+6c0fTQiX5pCjGGTc6 H7Fk2uwOfG9Z2dSGxP2SFdZmVsh0Zejr5kSmZyr+X19P+eQ0WhHM+6hnaRpeEReDX+grSlkNpH6x yVpC8poWmwdTnwh8PXazfNV1ig6cT79aKWNyce0QDud0ZEbHkhr2qwJJ/TSqSpDAlOM66OnoUcy8 PPlNi2NqYJH4I56gupCI+8TERnyiRWEaei//wNuIUISpKSYuho4Vd1E2f3sNjaVSXPQkCZAoP7+9 YSRLozuxxkGPjaM2c1mu4EOYEbV6jqVp6Frm/O1FuL455vYdWdYshPmzjvIfUXyIsqbl8DF0sFOB S+7NkRVubLsaBNZlKRp5n9y/rWMoCd6iDTH7l7DisDexBiYUaTqI0Sb/xHuSBvgvnc5W2uLSxYB9 C7dyJdQAA+PK9JrbgxrpH/U3q6m2J7nwbDYXns8hMvqdKKyejoychmYYG+RAHz10+Pi0NSomGsFr RCvkhCsiEuMVI31zfrQdy49Fx2k34BQgebRzEs4Bjdk+pCYREcYYG0PQwbn8edMOV6dfebS0D4fy T2Cu4EnkEURgjLFMzo1141iuEOCqmSjTu+ML+fNkQSZO74RdyGlcJ2xEr6sqJJW5NW4Z71rPpn/t +OhE9CR5cah6n42e5Rk3uQN2F93ofN6GeeNaYZPx86B2OtSgttqQCG1GKt5x3XclT94c5UXIZRSx n38Mr69rRKFc9pTI05TqNgMQQNH6+gQSOZdWDmWH6SAWd9Tj0PKdHPOLgrA3vMrVSIwT/FUhCfJg 44oD3AmDiHeBBFfozvYh9RLFEoBbLu/Mom7lAZXpJtGTdMDCYytuOy/gY1KZrg59aOjtTuedT7AI j6XK4Pk41JCB3ItDy7fx178RlGrUmb4dq5EVn+1pBImqcWPjFLwKu827CG+iY97zLvJ54m0zQ2sR BjOjQliZVkYAJUMvVUjkoTw4u5VVOyNo7jKc8pdcGPGkFutGNOSD4EnOF02E5EC+MSzobMeV1YPY ZjKIRd1Ki55k3vtWSSDx3+/CiEvFmZyqJ0kITuXc3TKZmcFN2V7hnjjdjG72mlU7oeu0wdRPICLi JPMcTlPYeQadsmBUm25IMtTomjamurrRN6ZA8Zq07NGNhrYyePQ3U+YfxtfYkpolTDntp4TE8rw7 E9Z4EFqjL8sqPsR53VUUuYtT1uYdF/SaJoFE8AC7Zi1l3yM5ZuXtKRF8Ab0WqtNNTR7OXcJRv2ii o61p7DiSDgHu8TFJG2J2zmC2MOU0j2DVzjtER0djWKk9EwfUy16eRFO7SuWzrgaypifJuvrOkpJL kGRJs31doSVIvq6+s2RvEiRZ0mxfV2idgICwpD/GfN3+pd6ygAYkT5IFjPStRZQg+dYWyAL9S5Bk ASN9axElSL61BbJA/xIkWcBI31pErSCJiYnh5ctXBLx+zdu3bxEOAOvq6WJkaIilZX6K2BZBpq/N +zxTVk+GpZ54uJEuXVfyb2I3vyQe5pK/OsuysbPYci8U/STniuW8PLOS0VN34xmqT5GGA5g7swOq 21G+tVEzuv90QRIbG8t9Ly98fHwR/i1cJibG6OjoEqNQEBUdTVxcHDo6OlgXsqZsmTIYGKS8lTE9 A8qw1BM3l/OLw0sc9g9AuZvEiNzi/lUfdg7uwar841k/rBL3F/Vj9O1WynPFPrsY0GkjVs5ujKjy ANdeU7jVbiX7+lVMz1CyRB2NIZErFFy7dp3g4GARDOuC1hS0Lohpzo8HwIUyz72fE/A6gGfPvDE0 NKRFi+bkMDHJQKVkQOqJ1M4Lix7mBI02baRvWUCE6Raddq7G/kIf2qucOb7l1p5e19uwd113imfg 6DJTUxpBEhERgcfly4SHR1CggBVVKlcWvcWH6ACMZLkIifQV/zOR5SF/zvLo6Ojj6XmPW7dvYW5u LoKScdOPmqknSl5nfsuRnG+1jN2Dq/Lh5ExaTQpl9IE5NLspnBc+JtrDwMyOtuMnMbpJUWTJ4ZEf Y6L9FGRzL1H34g8MJyFxTvY4c6wRJDdu3hRjkBIlilOmdGlxo9F13xVEKN5Su8hw/EOvcfn5YlHp xrI8dK28Hz1dQ+7c8eT2nTuULFGCqlWrYGSUEZuQ1E89YXNmNu3Gv6T/5l68Hj2c/37/C9eWeZGH BhAoHkF9x5Wl45h83IbxQhKcx8kzEpxkSjUnSAGS7JC9QG1I/P398bx7lzx58lKjejXx8NXJp5OI UoSKUFQv1J+QCB/OPZsh/m1n2Q6bXLXFbY7CntaLly7x5MlTatasgV2ZMhngTTVJPRHEUafuzLwq I7pET3a6deCTZEgh/zDiZxfM5l5iCskgkTyJkJPo89f1GzcIDAyiUsUK5MtvTtD7Bxgb5hE9iYVJ CcpbdcFALwdhUX6ERLwgl1FhVl2tSnh0EN2rHCafQS0OHzlKrly5aPhzA/S1XvVoknpCmU9t3JFw FLWGc2R+y093iImQzMNiwVkmmQkxyDU6bF1H79JJY5IKR9rTy6MVuzf3FM/XSDGJCjv/HDxEvnx5 qVWzJr4hHrwOu0su48KUyNMsRcL+e3OCjTcbi/fKWLahc6W9nD17Du/nz2nRvBn58uX7PJlpllA/ 9YTi5Ew6TQpl2Ka6nO6xDOOpm5nZJC9Pz/xDgE1tSpu/4ty8cUy+Ys/CvyfSMJcnK9sOYFdlZ9YP Lss9t8E4BfRkj+CBPFfTus9+qsxwY0j5+yz9Yxav/tjEyg7fb6I2tacbAZJixYpS1s6OO/4bxZ3y wgbnSgV7JpoyNiwAXdOPr2a+6fc/QqP80EWPn4pNwuvBA65evYa9fS1Kl9J2R7CaqSc4yZRu03ja w51N3UvzaLMDXddYM3PXFCpcnEDn2WfEdJ7mxeswcJITXSoqj0jI7+9i/MSVHPNRYF6+PS4LhvCj uLFZjteuyYxZcobnCjMqdp7EoqE/Zsm9q+p+SzWCxM6uDMWLFeWm3xri4mLJm6MMRXL/JPYVecAR w4dLiY3VIy7u09wjsbo5edfmOIc9vKhWrSrly5VTV0ap3DfWgEaQVK1ShYIFCygzC+joopo/Lezs egJX9UHP3Aod2aerFz3TvMi7unP2+j1+rFOH4sWLfeOhS92rqwGNIClRvDhlypTmZdgtohVh4vLW OlfNRFhiwoIQYEjtevjwEZevXKFxo4YULFhQXRmlct9YA2pDcuLkSXKY5KB2bXveRjwjJP4wVgHT Khjq51JrGAIgAihtWrcSVznSlTU0oDYkwoO0168Dsa9VE9NcxviHXic2LkY8wqmv+/nU4IGBgZw7 f4Hcuc35qW5d7ZbACYezSrcVz/km+CTxDO8GT2r8sY6haWUASNU2Z1nc/ThWCSftgm6wY912Dt0J Iko4BFauKb//2ZoKxiCePd7vJ7akb2JDgz6O9K2dl6Abu1iz+QS3AqNB35aOTpNpm8Wf16sNiY+P D55372FpmY+aNWoQIQ9GERuFqWEBtb4OV69dx8vLi5/q/kjRokXVqpNqIRGS07wxKcAvk5xoK64+ fdgz2Z3LUX5YC7lEtIWkqBdbnVy5a9eXQR3tseEJB92Wsi2iES5OvxKZeEC9HhE3NjBulYKu7tXx GrqK4F+dGdLoS+RK0U5t6a2tNiRyuRzPu//y8qU/RW1tKVdO+OVLvUv4ke/8hQvkz29J0yZN1KuU Vqn4s8AVmj/nOj1ZLBzsvreVobtiqKJ/ipC6AiQ3PklBMaj8XVwnHMZi0Ex6WxxhypSH2M8dRbPE MErFk7xdQ9+tBgxZ3IPKCbKEHGPW4IuUmDqVSjcSshgIB82FeqcoOLUvehunc61YPFjfSToktSER 9CT8unvhwkU+fPgg/nYj/IbzueuOpye3b9/BxMSEJo0bZUwsknBg3NmSQ/Pf0cbNAdOdY/jLtD12 l1YosxLVj0gxBYXB0fmMvFqE1uYXeFjKifFNVQNtFUj83eh8uyybhjVQyZ4kZBjYAL/PoMrtBEh+ 4NWJFUw8lJdhrl2pIExR/9vMgUcm/NDdAYefbb9A9qXPaT1j72sEidD1h/BwLl68JB6CNjMzpUiR IhSytkZP72OiPAEmYXp69uwZfn5+5MtnSYP69TAWkoZkxJWYVeBXgifPJqBxV9h6jeKzWvJ2Vnzq qjpC2odPU1CU4h6bRy7gH5owcUGXZHnnVSAJWEn33TkY9YknuU7ZuROwu5AQk+iTw6oMrfo60lpl 51HEk0MsdT2Kea/Usy9lhCq+RhsaQyIIJQAixBdv34Xy/Lm3GITmzZtHlDcmJpagoKDETUfC8xD7 WrWSQKT1wFRST1Q+vpCh2x5gWLM3bgPysic+v1mH0JRTUFgInuS2DU2jL/GqjhMjfk7FkxS9w/qx y3hYSTl1WMkfcsx9OdtifmHR2CYEq8QkKY5H/h+7Z0zn8Q8LGd80A3KyaK209DeQLkgSunv//r24 O+25jw/h4eEI+02Eq0CBAtjYFKJw4cLiVsYMv5LkJ/FgxeC95HScS4/yKglnCqaQgmKQEVudr2A3 YQZtwzcxavlbWrk40jClmKQUyF9dYOPyHZz77724urGt0pJu/Vp8XN2ImZU+Jr8BD/43cD0nwoSV jbAaak6/Yb9m+a2NWkGS4caXGsyUGpAgyZRmyVxCSZBkLntkSmkkSDKlWTKXUBIkmcsemVIaKfVE pjRL5hJK8iSZyx6ZUhoJkkxplswllARJ5rJHppRGgiRTmiVzCSVBkrnskSmlyRBIYuPknH82m0dB BymRpzn1ik1AVycLv5YhU5rq2wmlNSTvoxF0CtYAAAfQSURBVF5x6bkrDwP3iRujheOfJfM2x77w cMyNtX2L+LdTjNTzRw1oBYkqIAXMqtGizGIOPRjKy9Ab8aAMw9zYVtJ3FtdAuiF5H/USD5/FPHi9 l5J5f6GWzZ/kNinGu4jnXPN1wyvwbwqb/0j9Ys4SKNkRElUPIoy/RN6mWOYsTzXrftzwc+f1+3t4 +m8hQhFM5YK9xfftSR4l65KisSdJDkjC0IVDWm3Lb+Lgg8F4BewVE9vEoUyVlfGgKPOFPHRcx7Ze wrH/jL1OTkuaqCaxdTG5zRNGJGQbUKvbT88sq1VNjUI+J905GNOYP4TEO2qUT28RjSBJDRCh8wRI Dnj1x9N/M/LY8CQyZSwoEiRC4p0jk9oxJXwoe1xbJZ49Si8IadVTG5LkQaoQnCZcuYyL0LD4dCxM ahAUfpWDXgPFV68lXEJs4vPuQgZ6FAmSLwFDam2qDckBr354B58V44/ahUey6aby/Iyw7O1U6QDb HhrwIDgCCyMZQyrps+ZqWSJjlG/+dKzzmBNPxnM/YLf4qtf2Fbahq6NN6s7UIJHjfWwhk2YdxDNU njQPGnfYMnguq24+U6aaUEknIfc+xrzJruwR0nGW7kIby21stfiYFy35dNN3dFWurVKm6CzTcyrL xdQTqbWfLLfbGVcmuhzkxpsIMLGhieNkZnUoy7MNfWi/pAhDh79hg9sN3umXpM/sBQyvI2zADeLa Bldc1l/kqZAWtPUYVjnbcqx7H1xLKeUUp8hHXRlR/jwrd/misKrDmHkudCwrg6BrrJw6lfWXgsXP O1cNYOOzRomJeD4HnNqQLL1YmmJ5GtG8tDIn2vuoAPH/+noGeAYase1BICFRMUTHxtLVLi/1C0Wj iIkSywgvbBSu3Xe7iKC0KruaqtZ/fE62NO6nDMk7ITfayIvUnbGEEfUt8D86l8HTn9Jh3XYGVPTh 3o0YClcsipniLDNajOdh341s6h7D+u59WGs1EnfnFhT0P8SUgQs40SA1SA5StOMMljnWhptrGOO4 C5vp+5jbIjTN9hOM+c7rDi8sS1E+jz7/bR5KG7fCLPEYR0ERkiCaOc1j0s/6HBjflznhDmJWR9mu obR1Daf3/An0rBDGtkW3aeBci0vJITlcjn7znehZ3JsVg8axq5oL1ydW5uS0zgx/1JKVs/pShZu4 jx/NGgZ8GUhyGxejbYWNyHRz8OTNEdGIhnpmvFXYs/LOKz7IY8TP+lawopDxeSLkQeLfCS+K3nyz uTjttLRbQfVCAzIckjMz6+Ho/ydH3DrEz9E3WdbqTy62EwLcPFzbvpY1+67i/SoQ/1A55YTA98dr 9Oi4mUrLjjCqtlIk9QNXpZc42jiN9nshlkmARPBa/1u+h38evOCNXzDhKBMM24iQlEhMNvxI+Ht9 RVad7oTv4PYsK7aAcyOVAsofPsSnNJz/xJN8NLw4hoDhHHHLx9raE3gxYjcrOyhPTYttH2+Q8ZDs uNNOXNoKoDQqOYtdnp0Sjdym3CbO+BflysswqlvlpEWRUFZetkN4LaxwDap9l3+8BoqACPFJz2rH tXwNbMqeRITk7WjOzW+J8qSLEpIrHTcyx2QFrZaaMnnjWH618floXPsrdOm6nzqrd/NnVdEEHHWq x2hZ6tPNx9XNfdZ0+YMTzdaxwNQ95faTQGLKwqbDuNrOjVV9qmB27mMCvxQhEaHphF+fnuyovoyD g0UB46+k05hyukkGifi3NVurORE69RCuLZVa8Vrbk06nvsB0Izwku+yzkMdBhzGR5SU83ksIh8bl Me/5ucQM0Tvc9l/PkQfDE+MRQSghI1LQhweUt+osPjMR/tbuUkJyv+8iFrW2VjZlZIHJbVdajbnM z6rTzZzX9Fq/GvvLfWj/VyVW7xiM3YMN9P9jLYhL6BDRcP9UGYPbhAbke7yHYX+u5k7z1CA5RqXf XXDpUZ4Pp4TpzJOWKw/wy/3U2lf1JIhyvxn/F66/yDni0o+Jh+qm7kniPYule2e6rTOlnzjd6HHv 3H/kb1ng855EhKQR94Z2wSW0E0umdaN8jAeu/aexJ98XmG4EO6iCkmBk4XlIpCIE29z16V39NNvv /MaD139/wkDGASJOCKKy96j20low6g/c3T4d51VCgCfHpHBNeo2axEAh+As6iYuDC9t9oEiHXjS6 t5LL4jRRmqDrq5k4ejMe8bni2xguZHGqnuQhv3bNz43dF/HX/xh4ylJtXxWSn/DaPIL+bjcIN6qG Y488zHeTfRaShvKXnF0+lam7PQkKN04jcE3Jk/SkqPc/TB++kL/j878NrHaTWVe/wHSTYA8BlCu+ S3gU+A/RMWGERb0Ub5XL35kOFbex774Dt/z+lwSSjAVEOx8k1QaPBc3of6+72qnO1V7dqCpX6VEW ccVnKXHEYKyflxH1fJHpGiFsG1h0oRihkS/EKhIg3x7LR3uWcil/S9pWtSD45hacRm1GMUD9p9Xp gkQ59Xhz5r+pYgwiLHFH1PVN1MayS3YZHIN8e0VnZQn8j8xk8Jxj4hQsPJv5ofsgJv9ejwJqPstP NyTJQRHytFa06iY+Bzn11EnyIFmZqmSyawVJclAS2pammO+IEEBrSAR1RMe8Z/99B3FVUyJvM/HX YAO9j++/+b5Ulv1GkyGQZD+1Za8RS5BkL3una7QSJOlSW/aq9H9MIqy78+FAdAAAAABJRU5ErkJg gg== ------=_Part_513267_1414491071.1478284053203--