我正在尝试使用docker 1.13.1，kubernetes 1.9.6、1.10.1的alluxio 1.7.1
我按照https://www.alluxio.org/docs/1.7/en/Running-Alluxio-On-Docker.html上的说明创建了alluxio docker镜像
然后，我按照https://www.alluxio.org/docs/1.7/en/Running-Alluxio-On-Kubernetes.html指南在kubernetes上运行alluxio。我能够正确地启动alluxio master pod，但是当我尝试启动alluxio worker时，出现了Address在使用中的错误。我没有修改从alluxio git下载的yaml中的任何内容。我所做的唯一更改是针对Yamls中的alluxio docker镜像名称和api版本，以使k8s正确匹配。
我检查了k8s群集设置中使用的端口，甚至还检查了节点上是否正在使用。没有Alluxio希望任何其他进程使用的端口，但是我仍然收到使用中的地址错误。我无法理解我可以做些什么来进一步调试，或者我应该改变些什么才能使它正常工作。我的k8s群集设置上没有任何其他应用程序在运行。我尝试了单节点k8s群集设置和多节点k8s群集设置。我也尝试了k8s 1.9和1.10版本。
从Alluxio Worker方面肯定存在一些我无法调试的问题。
这是我从工作人员 pods 获得的日志:
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# kubectl logs po/alluxio-worker-knqt4
Formatting Alluxio Worker @ vm-sushil-scrum1-08062018-alluxio-1
2018-06-08 10:09:55,723 INFO Configuration - Configuration file /opt/alluxio/conf/alluxio-site.properties loaded.
2018-06-08 10:09:55,845 INFO Format - Formatting worker data folder: /alluxioworker/
2018-06-08 10:09:55,845 INFO Format - Formatting Data path for tier 0:/dev/shm/alluxioworker
2018-06-08 10:09:55,856 INFO Format - Formatting complete
2018-06-08 10:09:56,357 INFO Configuration - Configuration file /opt/alluxio/conf/alluxio-site.properties loaded.
2018-06-08 10:09:56,549 INFO TieredIdentityFactory - Initialized tiered identity TieredIdentity(node=10.194.11.7, rack=null)
2018-06-08 10:09:56,866 INFO BlockWorkerFactory - Creating alluxio.worker.block.BlockWorker
2018-06-08 10:09:56,866 INFO FileSystemWorkerFactory - Creating alluxio.worker.file.FileSystemWorker
2018-06-08 10:09:56,942 WARN StorageTier - Failed to verify memory capacity
2018-06-08 10:09:57,082 INFO log - Logging initialized @1160ms
2018-06-08 10:09:57,509 INFO AlluxioWorkerProcess - Domain socket data server is enabled at /opt/domain.
Exception in thread "main" java.lang.RuntimeException: java.lang.RuntimeException: io.netty.channel.unix.Errors$NativeIoException: bind(..) failed: Address in use
at alluxio.worker.AlluxioWorkerProcess.<init>(AlluxioWorkerProcess.java:164)
at alluxio.worker.WorkerProcess$Factory.create(WorkerProcess.java:45)
at alluxio.worker.WorkerProcess$Factory.create(WorkerProcess.java:37)
at alluxio.worker.AlluxioWorker.main(AlluxioWorker.java:56)
Caused by: java.lang.RuntimeException: io.netty.channel.unix.Errors$NativeIoException: bind(..) failed: Address in use
at alluxio.util.CommonUtils.createNewClassInstance(CommonUtils.java:224)
at alluxio.worker.DataServer$Factory.create(DataServer.java:45)
at alluxio.worker.AlluxioWorkerProcess.<init>(AlluxioWorkerProcess.java:159)
... 3 more
Caused by: io.netty.channel.unix.Errors$NativeIoException: bind(..) failed: Address in use
at io.netty.channel.unix.Errors.newIOException(Errors.java:117)
at io.netty.channel.unix.Socket.bind(Socket.java:259)
at io.netty.channel.epoll.EpollServerDomainSocketChannel.doBind(EpollServerDomainSocketChannel.java:75)
at io.netty.channel.AbstractChannel$AbstractUnsafe.bind(AbstractChannel.java:504)
at io.netty.channel.DefaultChannelPipeline$HeadContext.bind(DefaultChannelPipeline.java:1226)
at io.netty.channel.AbstractChannelHandlerContext.invokeBind(AbstractChannelHandlerContext.java:495)
at io.netty.channel.AbstractChannelHandlerContext.bind(AbstractChannelHandlerContext.java:480)
at io.netty.channel.DefaultChannelPipeline.bind(DefaultChannelPipeline.java:973)
at io.netty.channel.AbstractChannel.bind(AbstractChannel.java:213)
at io.netty.bootstrap.AbstractBootstrap$2.run(AbstractBootstrap.java:354)
at io.netty.util.concurrent.SingleThreadEventExecutor.runAllTasks(SingleThreadEventExecutor.java:399)
at io.netty.channel.epoll.EpollEventLoop.run(EpollEventLoop.java:305)
at io.netty.util.concurrent.SingleThreadEventExecutor$2.run(SingleThreadEventExecutor.java:131)
at java.lang.Thread.run(Thread.java:748)
-----------------------
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# kubectl get all
NAME DESIRED CURRENT READY UP-TO-DATE AVAILABLE NODE SELECTOR AGE
ds/alluxio-worker 1 1 0 1 0 <none> 42m
ds/alluxio-worker 1 1 0 1 0 <none> 42m
NAME DESIRED CURRENT AGE
statefulsets/alluxio-master 1 1 44m
NAME READY STATUS RESTARTS AGE
po/alluxio-master-0 1/1 Running 0 44m
po/alluxio-worker-knqt4 0/1 Error 12 42m
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
svc/alluxio-master ClusterIP None <none> 19998/TCP,19999/TCP 44m
svc/kubernetes ClusterIP 10.254.0.1 <none> 443/TCP 1h
---------------------
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# kubectl describe po/alluxio-worker-knqt4
Name: alluxio-worker-knqt4
Namespace: default
Node: vm-sushil-scrum1-08062018-alluxio-1/10.194.11.7
Start Time: Fri, 08 Jun 2018 10:09:05 +0000
Labels: app=alluxio
controller-revision-hash=3081903053
name=alluxio-worker
pod-template-generation=1
Annotations: <none>
Status: Running
IP: 10.194.11.7
Controlled By: DaemonSet/alluxio-worker
Containers:
alluxio-worker:
Container ID: docker://40a1eff2cd4dff79d9189d7cb0c4826a6b6e4871fbac65221e7cdd341240e358
Image: alluxio:1.7.1
Image ID: docker://sha256:b080715bd53efc783ee5f54e7f1c451556f93e7608e60e05b4615d32702801af
Ports: 29998/TCP, 29999/TCP, 29996/TCP
Command:
/entrypoint.sh
Args:
worker
State: Waiting
Reason: CrashLoopBackOff
Last State: Terminated
Reason: Error
Exit Code: 1
Started: Fri, 08 Jun 2018 11:01:37 +0000
Finished: Fri, 08 Jun 2018 11:02:02 +0000
Ready: False
Restart Count: 14
Limits:
cpu: 1
memory: 2G
Requests:
cpu: 500m
memory: 2G
Environment Variables from:
alluxio-config ConfigMap Optional: false
Environment:
ALLUXIO_WORKER_HOSTNAME: (v1:status.hostIP)
Mounts:
/dev/shm from alluxio-ramdisk (rw)
/opt/domain from alluxio-domain (rw)
/var/run/secrets/kubernetes.io/serviceaccount from default-token-7xlz7 (ro)
Conditions:
Type Status
Initialized True
Ready False
PodScheduled True
Volumes:
alluxio-ramdisk:
Type: EmptyDir (a temporary directory that shares a pod's lifetime)
Medium: Memory
alluxio-domain:
Type: HostPath (bare host directory volume)
Path: /tmp/domain
HostPathType: Directory
default-token-7xlz7:
Type: Secret (a volume populated by a Secret)
SecretName: default-token-7xlz7
Optional: false
QoS Class: Burstable
Node-Selectors: <none>
Tolerations: node.kubernetes.io/disk-pressure:NoSchedule
node.kubernetes.io/memory-pressure:NoSchedule
node.kubernetes.io/not-ready:NoExecute
node.kubernetes.io/unreachable:NoExecute
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal SuccessfulMountVolume 56m kubelet, vm-sushil-scrum1-08062018-alluxio-1 MountVolume.SetUp succeeded for volume "alluxio-domain"
Normal SuccessfulMountVolume 56m kubelet, vm-sushil-scrum1-08062018-alluxio-1 MountVolume.SetUp succeeded for volume "alluxio-ramdisk"
Normal SuccessfulMountVolume 56m kubelet, vm-sushil-scrum1-08062018-alluxio-1 MountVolume.SetUp succeeded for volume "default-token-7xlz7"
Normal Pulled 53m (x5 over 56m) kubelet, vm-sushil-scrum1-08062018-alluxio-1 Container image "alluxio:1.7.1" already present on machine
Normal Created 53m (x5 over 56m) kubelet, vm-sushil-scrum1-08062018-alluxio-1 Created container
Normal Started 53m (x5 over 56m) kubelet, vm-sushil-scrum1-08062018-alluxio-1 Started container
Warning BackOff 1m (x222 over 55m) kubelet, vm-sushil-scrum1-08062018-alluxio-1 Back-off restarting failed container
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# lsof -n -i :19999 | grep LISTEN
java 8949 root 29u IPv4 12518521 0t0 TCP *:dnp-sec (LISTEN)
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# lsof -n -i :19998 | grep LISTEN
java 8949 root 19u IPv4 12520458 0t0 TCP *:iec-104-sec (LISTEN)
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# lsof -n -i :29998 | grep LISTEN
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# lsof -n -i :29999 | grep LISTEN
[root@vm-sushil-scrum1-08062018-alluxio-1 kubernetes]# lsof -n -i :29996 | grep LISTEN
alluxio-worker容器始终重新启动，一次又一次由于相同的错误而失败。
请指导我该怎么做才能解决这个问题。
谢谢
最佳答案
问题是UNIX域套接字路径短路。我使用的是alluxio git中默认存在的内容。在默认的integration/kubernetes/conf/alluxio.properties.template中，ALLUXIO_WORKER_DATA_SERVER_DOMAIN_SOCKET_ADDRESS的地址不完整。 https://www.alluxio.org/docs/1.7/en/Running-Alluxio-On-Docker.html中对此进行了正确的说明，以允许使用unix域套接字在alluxio worker容器中进行短路读取。
仅由于缺少unix域套接字的完整路径，当为alluxio worker启用了短路读取时，alluxio worker无法出现在kubernetes中。
当我更正ALLUXIO_WORKER_DATA_SERVER_DOMAIN_SOCKET_ADDRESS =/opt/domain/d的Integration/kubernetes/conf/alluxio.properties中的路径时
然后事情开始正常运作。现在，一些测试也失败了，但从根本上来说，Alluxio的设置已正确启动。现在，我将调试一些测试失败的原因。
我已经在alluxio git中提交了此修复程序，以便他们将其合并到master分支中。
https://github.com/Alluxio/alluxio/pull/7376
关于sockets - 无法在Kubernetes上启动Alluxio，我们在Stack Overflow上找到一个类似的问题：https://stackoverflow.com/questions/50759565/