Create multiple stacks with the same efs

[lucienwang1009]

When I create two stacks with the same efs, it fails and rollbacks with the following error:
requested subnet for new mount target is not in the same VPC as existing mount targets (Service: AmazonElasticFileSystem; Status Code: 409; Error Code: MountTargetConflict; Request ID: 57f54cf9-cdf7-11e8-ade0-4fc9b2d0aee0)

I believe it is common to create the different stacks with the same efs, so can anyone please resolve this bug?

[mirocody]

Hi @lucienwang1009, can you provide minimum code example/setting to reproduce your issue?
Also, I am curious why you need to create different stack with same EFS, do you just need more work nodes, or new master nodes as well?

[lucienwang1009]

Hi, I have several deep learning tasks that use the same dataset. So I put the dataset onto my EFS and create several stacks with the same EFS ID to run different tasks. The template I use is exactly cfn-template/deeplearning.template with a specific EFSFileSystemId.

Following the instruction, I write the scripts for each task to create the stack, upload the training code and run the program automatically. Yes, I can do all these tasks in a common stack, but I don't want to mix them up in a shared stack. I'm just wondering why EFS can be mounted to different VMs but cannot be shared by different stacks.