So basically, I created a large windows server for development, and then I created a micro windows server for production. I set up everything how I wanted it on my development server, and then i unmounted the drives, and mounted them to my micro server.

Now I'm trying to get back into my large windows development server, and I'm getting the error.

Invalid value 'i-4896ce28' for instanceId. Instance does not have a volume attached at root (/dev/sda1)

this error pops up when I try to start my large windows server. I've remounted the drives to the large development server, and I still get this message.

I'm not really sure what to do, I've read other posts and everyone is giving these almost like command line arguments and talking about other tools, and I really have no clue what any of that means, or where I even have an option to enter any commands without be logged into a specific instance.

6 Answers 6


Make sure the EBS-drive has the name /dev/sda1 in the EC2-Console.

  • 2
    It was confusing to me how they put on the botton of the attach device dialog that windows devices should be xvdg-xvdh..I wasn't aware I could just type in /dev/sda1. Thanks alot.
    – Kyle
    Oct 29, 2011 at 2:56
  • I've been there to; wouldn't know it otherwise ;) Oct 29, 2011 at 6:46
  • In my case I had to enter "/dev/xvda" in the input field as "/dev/sda1" kept leading to the same error message. Using Debian Wheezy image.
    – Greendrake
    Oct 19, 2014 at 6:55
  • What a strange behavior.
    – David
    Feb 16, 2016 at 15:38

You can try /dev/xvda. When you attach a volume to the instance -> select the particular instance then enter /dev/xvda for the mount point.


Make sure all of your volumes are detached before re-attaching. In my case I had a volume attached to an earlier snapshot. Once i detached that I was allowed to attach my new volume.


Under volumes > 'Detach volume' and then 'Attach Volume' and in the device field type the correct path. For example you might need to type: /dev/sda

  • +1 this actually worked for me, It boot up without any problems.
    – Neo
    Mar 25, 2015 at 23:32

Make sure there is no leading space in '/dev/sda'

For some reason copying text out of the aws console often attaches a space in front of the string. A very annoying side effect that had me spinning my wheels for 15 minutes trying to get the primary drive on my production server remounted. I tried all of these solutions before figuring that out.


After trying all this I finally fixed it by using /dev/sda1

The error being it seemed you can only input /dev/sdf up to /dev/sdp

But it used /dev/sda1 correctly.

For the steps:

  1. Note down the device name in error while starting the EC2 instance.
  2. Detach the currently connected Volume with EC2
  3. Attach volume to EC2

    Try to attached same volume to EC2 instance. However make sure that you are not using default “Device” such as “/dev/sdf” as below while attaching volume to EC2 instance.

  4. Start EC2 instance

taken from linuxroutes

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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