0

I have a really old m3 server (Ubuntu 14.04) that has a 300GB mounted EBS volume: /dev/xvdf. I also have a new m5 server (Ubuntu 18.04) that uses the Nitro system with a premounted EBS volume: /dev/nvme1n1. I tried unmounting then detaching the old EBS volume so I could attach and mount to the new server. However, that doesn't work as the instance ID doesn't show up during the Attach Volume screen in AWS console. It doesn't show up even if the new server has been turned off. I assume it's because of incompatibilities between the underlying AWS system for the servers.

That being said, I need to copy over the data from the old EBS to the new EBS. I read in a different SO answer about making a 3rd instance where I can attach both. Then do a rsync between the two. I don't think that'll work because of the mounting problem listed above. And I think that means I can't use AWS CLI either.

Both servers are in the same AZ. It's about 250GB of data that needs to be transferred. I was going to do the naive approach of simply scp over ssh.

My Q is, is there a better way than scp?

risa_risa
  • 727
  • 1
  • 8
  • 20
  • rsync over ssh? – Evhz Aug 09 '20 at 22:13
  • yeah, that's what i am going to go with. the old EBS is always being written to, so I can at least start rsync some time before actual cutover and do it a few other times to get the newer/updated stuff. thanks! – risa_risa Aug 15 '20 at 18:45

2 Answers2

1

Why not to use rsync over ssh, something like:

rsync -avzhe ssh --progress /mnt/nvme1n1 user@192.0.0.100:/mnt/rpmpkgs
Evhz
  • 8,852
  • 9
  • 51
  • 69
1

The easiest is just use recursive scp. Its not naive at all. Its the correct tool for the job imho. No need to sync, use a 3rd instance, or mount commands. Just scp away!

Rodrigo Murillo
  • 13,080
  • 2
  • 29
  • 50