I am running dependency application in marathon, one is mongo DB and other is karesti voting app. I have deployed these applications in marathon within a docker container. First I deployed mongo db and later the voting app.
我在marathon中运行dependency application，一个是mongo DB，另一个是karesti voting app。我已经在marathon容器中部署了这些应用。首先我部署了mongo db，然后是投票应用。
mongo db docker image is working fine in marathon. the issue is with voting app.... voting app is switching between the running mode and suspend mode continuously. Due to this the port number is also getting changed in each deployment and a new container is creating the docker for the same image.
mongo db docker图像在马拉松比赛中运行良好。的问题是投票应用....投票app在运行模式和暂停模式之间不断切换。由于这个原因，在每次部署中端口号也被更改，一个新的容器正在为相同的映像创建docker。
Here I have mesos DNS for service discovery.
这里有用于服务发现的mesos DNS。
Error from marathon:
从马拉松错误:
Docker container run error: Container exited on error: exited with status 1
Docker容器运行错误:容器退出错误:退出状态1。
Error from voting app:
从投票程序错误:
Received killTask for task cmvoting.dff585d5-2bc5-11e6-a4c0-0242efa3f3aa
收到任务cmvoting.dff585d5-2bc5-11e6-a4c0-0242efa3f3aa的killTask
Log message from mesos slave:
I0607 07:24:04.297161 1957 slave.cpp:3243] Handling status update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 from executor(1)@127.0.0.1:50983
I0607 07:24:04.299815 1957 status_update_manager.cpp:320] Received status update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:04.300282 1957 status_update_manager.cpp:824] Checkpointing UPDATE for status update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:04.391424 1957 slave.cpp:3641] Forwarding the update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 to master@172.31.29.198:5050
I0607 07:24:04.391826 1957 slave.cpp:3551] Sending acknowledgement for status update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 to executor(1)@127.0.0.1:50983
I0607 07:24:04.421507 1951 status_update_manager.cpp:392] Received status update acknowledgement (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:04.421766 1951 status_update_manager.cpp:824] Checkpointing ACK for status update TASK_RUNNING (UUID: ed039d5b-ef25-45e5-962f-367feaa33ae5) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:18.299789 1956 slave.cpp:3243] Handling status update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 from executor(1)@127.0.0.1:50983
I0607 07:24:18.321115 1953 status_update_manager.cpp:320] Received status update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:18.321229 1953 status_update_manager.cpp:824] Checkpointing UPDATE for status update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:18.338953 1956 slave.cpp:3641] Forwarding the update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 to master@172.31.29.198:5050
I0607 07:24:18.339341 1956 slave.cpp:3551] Sending acknowledgement for status update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 to executor(1)@127.0.0.1:50983
I0607 07:24:18.350075 1954 status_update_manager.cpp:392] Received status update acknowledgement (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:18.350272 1954 status_update_manager.cpp:824] Checkpointing ACK for status update TASK_FAILED (UUID: e20d07aa-88e1-497b-9c51-79252c279f27) for task voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3 of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001
I0607 07:24:19.301136 1951 slave.cpp:3769] executor(1)@127.0.0.1:50983 exited
I0607 07:24:19.327797 1953 docker.cpp:2011] Executor for container 'd4447d21-23da-45db-a9aa-f6b2d75e2db7' has exited
I0607 07:24:19.327910 1953 docker.cpp:1731] Destroying container 'd4447d21-23da-45db-a9aa-f6b2d75e2db7'
I0607 07:24:19.328034 1953 docker.cpp:1859] Running docker stop on container 'd4447d21-23da-45db-a9aa-f6b2d75e2db7'
I0607 07:24:19.328732 1958 slave.cpp:4120] Executor 'voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3' of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 exited with status 0
I0607 07:24:19.328843 1958 slave.cpp:4224] Cleaning up executor 'voting.cfc5c791-2c80-11e6-b454-0242f9dcd7c3' of framework 4925c40b-fb4b-4ec7-91fc-3ffc16e1bac8-0001 at executor(1)@127.0.0.1:50983
I am not able to identify where is the issue, why the voting app is been failed in slave.
我无法确定问题出在哪里，为什么投票应用在slave中失败。
If I run the voting from bare ubuntu machine, I am able to access the application, which has linked to the database and everything is working fine. Which shows that no issue with the app, no issue with the docker image.
如果我在裸机上运行投票，我就可以访问这个连接到数据库的应用程序，一切都运行得很好。这说明这个应用没有问题，docker图像也没有问题。
My suspect is that, something is happening in mesos slave which is not able to handle or might not be able to connect to the database.
我的猜测是，在mesos slave中发生了一些事情，它无法处理或可能无法连接到数据库。
my mongo db listen to the the port 27017 and voting app listen to port 9000. I can see the only the port 27017 is being used when i do the netstat -anp | grep 27017
我的mongo db监听端口27017，投票应用程序监听端口9000。我可以看到，在执行netstat -anp | grep 27017时，唯一使用的端口是27017
The command used to run the slave is
用于运行从服务器的命令是
sudo ./bin/mesos-slave.sh --master=zk://172.31.36.222:2181/home/ubuntu/mesos --executor_registration_timeout=5mins --containerizers="docker,mesos"
The command used to run the marathon is
跑马拉松的命令是
./bin/start --master zk://172.31.36.222:2181/home/ubuntu/mesos --zk zk://172.31.36.222:2181/home/ubuntu/marathon-1.1.1 --task_launch_timeout 600000
Please let me know if you need any more info or clarity on the problem
如果你需要更多的信息或清晰的问题，请告诉我
Thanks in advance for your kind help....
提前感谢你的帮助....
2 个解决方案
#1
1
One part of the problem could be that your slave is binding to the 127.0.0.1 ip address. Please try to set a non-local address on your slave like this:
问题的一部分可能是您的从服务器绑定到127.0.0.1 ip地址。请试着在您的奴隶上设置一个非本地地址如下:
echo 172.31.36.223 | sudo tee /etc/mesos-slave/ip
sudo cp /etc/mesos-slave/ip /etc/mesos-slave/hostname
Replace 172.31.36.223 with the real ip address of your slave.
将172.31.36.223替换为您的奴隶的实际ip地址。
Another part could probably be that the voting app can't connect to MongoDB because the port number is different. You should also post your Marathon app definitions to be able to verify this.
另一部分可能是投票应用程序无法连接到MongoDB，因为端口号不同。你还应该发布你的马拉松应用程序定义来验证这一点。
If you just have one slave, you could run the both apps in HOST networking mode, so that you can define the port numbers statically. If you're using bridged mode, you need to employ some kind of service discovery mechanism so that the voting app can find the MongoDB server.
如果您只有一个从属，您可以在主机网络模式下运行这两个应用程序，以便您可以静态地定义端口号。如果使用的是桥接模式，则需要使用某种服务发现机制，以便投票应用程序能够找到MongoDB服务器。
#2
0
I thanks for the response I have got...
谢谢你的回复。
I found the solution to this problem
我找到了解决这个问题的办法
The DNS looks for the right IP address from the files /etc/resolv.conf and /etc/hosts
DNS从文件/etc/resolv中查找正确的IP地址配置和/ etc / hosts
I removed the localhost IP address from /etc/hosts file, where DNS was trying to resolve from 127.0.0.1 which is unknown to external world. I provide my VM private IP address which same has been provided in the DNS conf file
我从/etc/hosts文件中删除了localhost的IP地址，DNS试图从127.0.0.1解决，这是外部世界所不知道的。我提供了我的VM私有IP地址，DNS conf文件中也提供了这个地址
#1
1
One part of the problem could be that your slave is binding to the 127.0.0.1 ip address. Please try to set a non-local address on your slave like this:
问题的一部分可能是您的从服务器绑定到127.0.0.1 ip地址。请试着在您的奴隶上设置一个非本地地址如下:
echo 172.31.36.223 | sudo tee /etc/mesos-slave/ip
sudo cp /etc/mesos-slave/ip /etc/mesos-slave/hostname
Replace 172.31.36.223 with the real ip address of your slave.
将172.31.36.223替换为您的奴隶的实际ip地址。
Another part could probably be that the voting app can't connect to MongoDB because the port number is different. You should also post your Marathon app definitions to be able to verify this.
另一部分可能是投票应用程序无法连接到MongoDB，因为端口号不同。你还应该发布你的马拉松应用程序定义来验证这一点。
If you just have one slave, you could run the both apps in HOST networking mode, so that you can define the port numbers statically. If you're using bridged mode, you need to employ some kind of service discovery mechanism so that the voting app can find the MongoDB server.
如果您只有一个从属，您可以在主机网络模式下运行这两个应用程序，以便您可以静态地定义端口号。如果使用的是桥接模式，则需要使用某种服务发现机制，以便投票应用程序能够找到MongoDB服务器。
#2
0
I thanks for the response I have got...
谢谢你的回复。
I found the solution to this problem
我找到了解决这个问题的办法
The DNS looks for the right IP address from the files /etc/resolv.conf and /etc/hosts
DNS从文件/etc/resolv中查找正确的IP地址配置和/ etc / hosts
I removed the localhost IP address from /etc/hosts file, where DNS was trying to resolve from 127.0.0.1 which is unknown to external world. I provide my VM private IP address which same has been provided in the DNS conf file
我从/etc/hosts文件中删除了localhost的IP地址，DNS试图从127.0.0.1解决，这是外部世界所不知道的。我提供了我的VM私有IP地址，DNS conf文件中也提供了这个地址