2.centos7 安装Mesos和marathon

master节点

节点名称	IP地址
master1	192.168.161.128

由于硬件资源有限,这里仅安装一个master节点,会存在单点问题.在实际生产中,尽量不要这样操作.

1.安装zookeeper

https://www.cnblogs.com/gytangyao/p/10172265.html

2.安装Mesos

1)添加mesos yum源

sudo rpm -Uvh http://repos.mesosphere.io/el/7/noarch/RPMS/mesosphere-el-repo-7-1.noarch.rpm

yum -y install mesos marathon

2)更改mesos的zk配置

vi /etc/mesos/zk 

写入zookeeper节点的地址,例如:zk://192.168.1.110:2181,192.168.1.111:2181,192.168.1.112:2181/mesos

zk://192.168.161.128:2181/mesos

按需更改quorum,如果zookeeper以单机模式运行则不用修改

设置 --quorum=2 的意思是：至少要有 2 台 master 节点在线，否则集群无法起来，所以至少要两台 master 一起启动。

vi /etc/mesos-master/quorum

3)配置mesos-master和marathon 的hostname和ip

hostname直接配置的本机ip，是为了防止出现dns无法解析的问题；

echo "192.168.161.128" > /etc/mesos-master/ip

echo "192.168.161.128" > /etc/mesos-master/hostname

mkdir -p /etc/marathon/conf/

echo "192.168.161.128" > /etc/marathon/conf/hostname

echo "192.168.161.128 master1" >> /etc/hosts

4)给marathon填加mesos的地址参数

vim /etc/default/marathon

MARATHON_MASTER="zk://192.168.161.128:2181/mesos"
MARATHON_ZK="zk://192.168.161.128:2181/marathon"
MARATHON_MESOS_USER="root"

5)启动服务

systemctl start  zookeeper && systemctl start mesos-master && systemctl start marathon && systemctl disable mesos-slave

6)其他配置

Marathon WebUI默认的端口是8080，修改端口的方法：

vi /etc/default/marathon

export HTTP_PORT=8181

export MARATHON_HTTP_PORT=8181

6)错误排错

Required option 'master' not found    可使用 journalctl -u marathon 查看日志

# journalctl -u marathon
-- Logs begin at Thu 2018-06-14 15:55:28 CST, end at Thu 2018-06-14 19:52:34 CST. --
Jun 14 15:55:35 dev-master-01 systemd[1]: marathon.service holdoff time over, scheduling restart.
Jun 14 15:55:35 dev-master-01 systemd[1]: Starting Scheduler for Apache Mesos...
Jun 14 15:55:35 dev-master-01 systemd[1]: Started Scheduler for Apache Mesos.
Jun 14 15:55:35 dev-master-01 marathon[7966]: No start hook file found ($HOOK_MARATHON_START). Proceeding with the start script.
Jun 14 15:55:36 dev-master-01 marathon[7966]: [scallop] Error: Required option 'master' not found
Jun 14 15:55:36 dev-master-01 systemd[1]: marathon.service: main process exited, code=exited, status=1/FAILURE
Jun 14 15:55:36 dev-master-01 systemd[1]: Unit marathon.service entered failed state.
Jun 14 15:55:36 dev-master-01 systemd[1]: marathon.service failed.
Jun 14 15:56:36 dev-master-01 systemd[1]: marathon.service holdoff time over, scheduling restart.
Jun 14 15:56:36 dev-master-01 systemd[1]: Starting Scheduler for Apache Mesos...
Jun 14 15:56:36 dev-master-01 systemd[1]: Started Scheduler for Apache Mesos.
Jun 14 15:56:36 dev-master-01 marathon[8046]: No start hook file found ($HOOK_MARATHON_START). Proceeding with the start script.
Jun 14 15:56:37 dev-master-01 marathon[8046]: [scallop] Error: Required option 'master' not found
Jun 14 15:56:37 dev-master-01 systemd[1]: marathon.service: main process exited, code=exited, status=1/FAILURE
Jun 14 15:56:37 dev-master-01 systemd[1]: Unit marathon.service entered failed state.
Jun 14 15:56:37 dev-master-01 systemd[1]: marathon.service failed.

如提示:java.lang.IllegalStateException: Multiple election participants have the same ID: master1:8080. This is not allowed.

则可能是多次执行了设置ip或者hostname,需要检查第三步的配置是否正确

12月 25 12:28:08 master1 marathon[8724]: java.lang.IllegalStateException: Multiple election participants have the same ID: master1:8080. This is not allowed.
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.core.election.CuratorElectionStream$CuratorEventEmitter.emitLeader(CuratorElectionStream.scala:157)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.core.election.CuratorElectionStream$CuratorEventEmitter.$anonfun$longPollLeaderChange$1(CuratorElectionStream.scala:117)
12月 25 12:28:08 master1 marathon[8724]: at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
12月 25 12:28:08 master1 marathon[8724]: at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
12月 25 12:28:08 master1 marathon[8724]: at java.lang.Thread.run(Thread.java:748)
12月 25 12:28:08 master1 marathon[8724]: [2018-12-25 12:28:08,788] INFO  Closing leader latch (mesosphere.marathon.core.election.CuratorElectionStream$:pool-1-thread-1)
12月 25 12:28:08 master1 marathon[8724]: [2018-12-25 12:28:08,812] INFO  Shutting down services (mesosphere.marathon.MarathonApp:shutdownHook3)
12月 25 12:28:08 master1 marathon[8724]: [2018-12-25 12:28:08,819] INFO  Shutdown triggered (mesosphere.marathon.MarathonSchedulerService:shutdownHook3)
12月 25 12:28:08 master1 marathon[8724]: [2018-12-25 12:28:08,819] ERROR abdicateLeadership was called (mesosphere.marathon.core.election.ElectionServiceImpl:shutdownHook3)
12月 25 12:28:08 master1 marathon[8724]: mesosphere.marathon.Exception: abdicateLeadership
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.core.election.ElectionServiceImpl.abdicateLeadership(ElectionService.scala:151)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonSchedulerService.triggerShutdown(MarathonSchedulerService.scala:169)
12月 25 12:28:08 master1 marathon[8724]: at com.google.common.util.concurrent.AbstractExecutionThreadService$1.doStop(AbstractExecutionThreadService.java:95)
12月 25 12:28:08 master1 marathon[8724]: at com.google.common.util.concurrent.AbstractService.stopAsync(AbstractService.java:242)
12月 25 12:28:08 master1 marathon[8724]: at com.google.common.util.concurrent.AbstractExecutionThreadService.stopAsync(AbstractExecutionThreadService.java:214)
12月 25 12:28:08 master1 marathon[8724]: at com.google.common.util.concurrent.ServiceManager.stopAsync(ServiceManager.java:331)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.$anonfun$shutdown$1(Main.scala:103)
12月 25 12:28:08 master1 marathon[8724]: at scala.Option.foreach(Option.scala:257)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.shutdown(Main.scala:103)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.$anonfun$shutdownAndWait$1(Main.scala:108)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.$anonfun$shutdownAndWait$1$adapted(Main.scala:107)
12月 25 12:28:08 master1 marathon[8724]: at scala.Option.foreach(Option.scala:257)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.shutdownAndWait(Main.scala:107)
12月 25 12:28:08 master1 marathon[8724]: at mesosphere.marathon.MarathonApp.$anonfun$start$1(Main.scala:80)
12月 25 12:28:08 master1 marathon[8724]: at scala.sys.ShutdownHookThread$$anon$1.run(ShutdownHookThread.scala:34)
12月 25 12:28:08 master1 marathon[8724]: [2018-12-25 12:28:08,822] INFO  backgroundOperationsLoop exiting (org.apache.curator.framework.imps.CuratorFrameworkImpl:Curator-Framework-0)
12月 25 12:28:10 master1 systemd[1]: marathon.service: main process exited, code=exited, status=137/n/a
12月 25 12:28:10 master1 systemd[1]: Unit marathon.service entered failed state.
12月 25 12:28:10 master1 systemd[1]: marathon.service failed.

Slave节点

节点名称	IP地址
slave1	192.168.161.200
slave2	192.168.161.201
slave3	192.168.161.202

1)添加 mesos的yum源

sudo rpm -Uvh http://repos.mesosphere.io/el/7/noarch/RPMS/mesosphere-el-repo-7-1.noarch.rpm

2)安装mesos

yum -y install mesos

3)配置master信息

master信息需要和前面安装的master节点对应

vim /etc/mesos/zk

zk://192.168.161.128:2181/mesos

4）mesos-slave配置ip&hostname

以下需要按照情况各自配置

slave1:

echo "192.168.161.200" > /etc/mesos-slave/ip
echo "192.168.161.200" > /etc/mesos-slave/hostname

#修改hosts
echo "192.168.161.200 slave1" >> /etc/hosts

slave2:

echo "192.168.161.201" > /etc/mesos-slave/ip
echo "192.168.161.201" > /etc/mesos-slave/hostname

#修改hosts
echo "192.168.161.201 slave2" >> /etc/hosts

slave3:

echo "192.168.161.202" > /etc/mesos-slave/ip
echo "192.168.161.202" > /etc/mesos-slave/hostname

#修改hosts
echo "192.168.161.202 slave3" >> /etc/hosts

5)启动mesos-slave

systemctl start  mesos-slave  && systemctl enable  mesos-slave && systemctl disable mesos-master

6)错误排查

如遇到slave连接不上master的情况,如果没有对mesos的日志存放目录进行特殊修改

请在 /var/log/mesos/  下查看日志

日志信息如下:

[root@slave2 ~]# cd /var/log/mesos/
[root@slave2 mesos]# ls
mesos-slave.ERROR                                               mesos-slave.slave2.invalid-user.log.INFO.20181225-141244.8972  mesos-slave.slave2.invalid-user.log.INFO.20181225-181648.8959     mesos-slave.slave2.invalid-user.log.WARNING.20181225-180358.8998
mesos-slave.INFO                                                mesos-slave.slave2.invalid-user.log.INFO.20181225-141514.8965  mesos-slave.slave2.invalid-user.log.WARNING.20181225-115141.9449  mesos-slave.slave2.invalid-user.log.WARNING.20181225-180823.8952
mesos-slave.slave2.invalid-user.log.ERROR.20181225-145851.8960  mesos-slave.slave2.invalid-user.log.INFO.20181225-141841.8986  mesos-slave.slave2.invalid-user.log.WARNING.20181225-115709.8954  mesos-slave.slave2.invalid-user.log.WARNING.20181225-181649.8959
mesos-slave.slave2.invalid-user.log.ERROR.20181225-155517.8987  mesos-slave.slave2.invalid-user.log.INFO.20181225-145406.8960  mesos-slave.slave2.invalid-user.log.WARNING.20181225-141254.8972  mesos-slave.WARNING
mesos-slave.slave2.invalid-user.log.ERROR.20181225-180825.8952  mesos-slave.slave2.invalid-user.log.INFO.20181225-155333.8987  mesos-slave.slave2.invalid-user.log.WARNING.20181225-141852.8986
mesos-slave.slave2.invalid-user.log.INFO.20181225-115131.9449   mesos-slave.slave2.invalid-user.log.INFO.20181225-180357.8998  mesos-slave.slave2.invalid-user.log.WARNING.20181225-145851.8960
mesos-slave.slave2.invalid-user.log.INFO.20181225-115659.8954   mesos-slave.slave2.invalid-user.log.INFO.20181225-180823.8952  mesos-slave.slave2.invalid-user.log.WARNING.20181225-155334.8987
[root@slave2 mesos]#