Got it. I have build mesos from source so I need to initialize log-dir manually for once. And now things are working.Thank you
-----Original Message----- From: "Badal Naik" <[email protected]> Sent: 09/10/2015 06:38 PM To: "[email protected]" <[email protected]> Subject: Re: After restarting cluster task disappeared No i have not deleted anything. i am just restarting physical nodes and then i am not able to check old completed tasks.And after restarting many times marathon and mess just stay out of sync and when i restart marathon service both get synced. On 09-Oct-2015, at 6:32 pm, haosdent <[email protected]> wrote: For #1, do you delete something in your work_dir or zookeeper? For #3, is this zookeeper issue related to yours http://stackoverflow.com/questions/15842553/zookeeper-network-ensemble-does-not-start-appropiately ? On Fri, Oct 9, 2015 at 8:30 PM, craig w <[email protected]> wrote: I'm not sure about #3. I have seen things go awry when restarting the whole cluster. When doing an upgrade from mesos 0.23.0 to 0.24.1, I restarted all of the mesos-masters. Waited a few moments for a leader to be elected, then restarted the slaves. When I went back to look at Marathon all of the tasks were being redeployed, as though they had all been killed off for some reason. That wasn't what I expected to happen since the upgrade was suppose to be as simple as install and restart. Perhaps you're experiencing a similar issue? On Fri, Oct 9, 2015 at 8:25 AM, Badal Naik <[email protected]> wrote: Any idea about #1 ? Any one has experienced #3 ? On 09-Oct-2015, at 5:53 pm, craig w <[email protected]> wrote: With regards to item #2, I saw the same issue. it's been fixed in mesos 0.25 (release candidates are out now), see https://issues.apache.org/jira/browse/MESOS-3282. On Fri, Oct 9, 2015 at 8:16 AM, Badal Naik <[email protected]> wrote: Hello Mesos-Users, I have set up 3 node mess cluster with ubuntu 14.04. i have started zookeeper,Mesos and marathon. Every thing working fine expect three things. 1) When i restart the whole cluster mesos does not show completed tasks. is it expected behaviour? if not what i should do? 2) in mesos web ui i’m not able to see staged/started/finished/killed/failed/lost task numbers even when tasks are running. 3) Every zookeeper instance throws this exception regularly: 2015-10-09 17:27:26,302 [myid:3] - WARN [SendWorker:1:QuorumCnxManager$SendWorker@679] - Interrupted while waiting for message on queue java.lang.InterruptedException at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.reportInterruptAfterWait(AbstractQueuedSynchronizer.java:2014) at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.awaitNanos(AbstractQueuedSynchronizer.java:2088) at java.util.concurrent.ArrayBlockingQueue.poll(ArrayBlockingQueue.java:418) at org.apache.zookeeper.server.quorum.QuorumCnxManager.pollSendQueue(QuorumCnxManager.java:831) at org.apache.zookeeper.server.quorum.QuorumCnxManager.access$500(QuorumCnxManager.java:62) at org.apache.zookeeper.server.quorum.QuorumCnxManager$SendWorker.run(QuorumCnxManager.java:667) Here is my Mesos-master configuration: mesos master --ip=10.1.0.72 --work_dir=/var/lib/mesos-master --zk=file:///etc/mesos/conf/zk --quorum=file:///etc/mesos/conf/quorum Where zk=zk://zoo.service.consul:2181/mesos quorum=2 Mesos-Slave Configuration: mesos slave --work_dir=/var/lib/mesos-slave --ip=10.1.0.72 --hostname=10.1.0.72 --strict=false --master=file:///etc/mesos/conf/master FrameworkInfo.checkpoint=True Marathon Configuration: java -jar /opt/marathon.jar --master zk://zoo.service.consul:2181/mesos --zk zk://zoo.service.consul:2181/marathon --ha --hostname 10.1.0.72 --checkpoint Zookeeper configs with java version "1.8.0_45": dataDir=/var/lib/zookeeper clientPort=2181 tickTime=2000 initLimit=10 syncLimit=20 autopurge.purgeInterval=0 zookeeper.connection.timeout.ms=6000 server.1=10.1.0.70:2888:3888 server.2=10.1.0.71:2888:3888 server.3=10.1.0.72:2888:3888 And different myid has been given. Can Anyone Help!!! -- https://github.com/mindscratchhttps://www.google.com/+CraigWickesserhttps://twitter.com/mind_scratchhttps://twitter.com/craig_links -- https://github.com/mindscratchhttps://www.google.com/+CraigWickesserhttps://twitter.com/mind_scratchhttps://twitter.com/craig_links -- Best Regards, Haosdent Huang
