My team has the task to build a plan in case the AZ that has the NFS instance fails. This NFS instance shares files with others machines (NFS clients) in others AZ (we use North of California).

We have to propose ideas. One we have is that, when the NFS instance AZ fails, disconnect it's hard drive and attach it to another instance in another AZ (all this using lambda) but with this we also have to reconfigure the fstab in the NFS clients.

Any ideas?


Instead of a server in an AZ use the AWS Elastic File System service. EFS supports NFS, and your files are stored in multiple AZs.

If an AZ goes down you can't address any resources in that AZ, which cuts out your lambda / reconnect idea. EBS volumes are in a single AZ. You could get around this with volume snapshots, which are stored in S3 multi-AZ, but then you'd have to restore the whole server. My understanding of Windows in AWS is limited, if the data is on the OS drive you'd have to sysprep it, if it's a data drive you could attach it to another server.

EFS is likely the best, simplest, and cheapest solution.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.