从 EC2 实例内调用 Spark-EC2：与主机的 ssh 连接被拒绝

为了运行 Amplab 的训练练习，我创建了一个密钥对us-east-1，已经安装了训练脚本（git clone git://github.com/amplab/training-scripts.git -b ampcamp4）并创建了环境。变量 AWS_ACCESS_KEY_ID 和 AWS_SECRET_ACCESS_KEY 按照中的说明进行操作http://ampcamp.berkeley.edu/big-data-mini-course/launching-a-bdas-cluster-on-ec2.html

正在运行

 ./spark-ec2 -i ~/.ssh/myspark.pem -r us-east-1  -k myspark --copy launch try1

生成以下消息：

 johndoe@ip-some-instance:~/projects/spark/training-scripts$ ./spark-ec2 -i ~/.ssh/myspark.pem -r us-east-1  -k myspark --copy launch try1
 Setting up security groups...
 Searching for existing cluster try1...
 Latest Spark AMI: ami-19474270
 Launching instances...
 Launched 5 slaves in us-east-1b, regid = r-0c5e5ee3
 Launched master in us-east-1b, regid = r-316060de
 Waiting for instances to start up...
 Waiting 120 more seconds...
 Copying SSH key /home/johndoe/.ssh/myspark.pem to master...
 ssh: connect to host ec2-54-90-57-174.compute-1.amazonaws.com port 22: Connection refused
 Error connecting to host Command 'ssh -t -o StrictHostKeyChecking=no -i /home/johndoe/.ssh/myspark.pem [email protected] 'mkdir -p ~/.ssh'' returned  non-zero exit status 255, sleeping 30
 ssh: connect to host ec2-54-90-57-174.compute-1.amazonaws.com port 22: Connection refused
 Error connecting to host Command 'ssh -t -o StrictHostKeyChecking=no -i /home/johndoe/.ssh/myspark.pem [email protected] 'mkdir -p ~/.ssh'' returned non-zero exit status 255, sleeping 30
 ...
 ...
 subprocess.CalledProcessError: Command 'ssh -t -o StrictHostKeyChecking=no -i /home/johndoe/.ssh/myspark.pem [email protected] '/root/spark/bin/stop-all.sh'' returned non-zero exit status 127

where [email protected]是用户和主实例。我试过了-u ec2-user并不断增加-w一直到600，但得到同样的错误。

我可以看到主实例和从实例us-east-1当我登录到AWS控制台时，我实际上可以从“本地”ssh到Master实例ip-some-instance shell.

我的理解是，spark-ec2 脚本负责定义主/从安全组（监听哪些端口等），我不必调整这些设置。这就是说，主人和奴隶都听22号帖子（Port:22, Protocol:tcp, Source:0.0.0.0/0在 ampcamp3-slaves/masters 秒。组）。

我在这里不知所措，在我将所有研发资金花在 EC2 实例上之前，希望能得到任何指点……谢谢。

这很可能是由于 SSH 在实例上启动需要很长时间，导致 120 秒超时在计算机登录之前到期。你应该能够运行

./spark-ec2 -i ~/.ssh/myspark.pem -r us-east-1  -k myspark --copy launch --resume try1

（与--resume标志）从中断处继续，无需重新启动新实例。这个问题将在 Spark 1.2.0 中得到解决，我们有一个新的机制可以智能地检查 SSH 状态，而不是依赖固定的超时。我们还通过构建新的 AMI 来解决 SSH 启动长时间延迟的根本原因。