Why does attaching a non-root EBS drive created from a snapshot not maintain the partitions?

Question

When I create a snapshot of an auxiliary RHEL8 EC2 root volume, I can then create a new EBS device from the snapshot. I have a need to attach the new available EBS snapshot as a secondary drive on my running RHEL8 EC2 instance. When I attach the EBS drive created from the snapshot, I'm seeing the root partition is no longer present on the disk. 
Here is a look at the lsblk configuration after attaching the secondary disk:

```
[root@ip-172-31-76-220 ec2-user]# lsblk
NAME        MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
nvme0n1     259:0    0   10G  0 disk
├─nvme0n1p1 259:2    0  200M  0 part /boot/efi
├─nvme0n1p2 259:3    0  512M  0 part /boot
└─nvme0n1p3 259:4    0  9.3G  0 part /
nvme1n1     259:1    0   10G  0 disk
```
Please note, /dev/nvme1n1 disk has no partitions, when this drive was created from a RHEl8 root device. I'm expecting at least one partition with the root OS present.

If I create a new auxiliary RHEL8 instance, stop the instance and detach the drive, I can then attach the available drive to my running EC2 instance and am able to see the OS partitions as expected. In the following lsblk config, the /dev/nvme2n1 disk retains the OS disk partitions.
```
[root@ip-172-31-76-220 ec2-user]# lsblk
NAME        MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
nvme0n1     259:0    0   10G  0 disk
├─nvme0n1p1 259:2    0  200M  0 part /boot/efi
├─nvme0n1p2 259:3    0  512M  0 part /boot
└─nvme0n1p3 259:4    0  9.3G  0 part /
nvme1n1     259:1    0   10G  0 disk
nvme2n1     259:6    0   10G  0 disk
├─nvme2n1p1 259:7    0  200M  0 part
├─nvme2n1p2 259:8    0  512M  0 part
└─nvme2n1p3 259:9    0  9.3G  0 part
```

adam gantt · Accepted Answer

Dag nab it, thanks for the response. After taking a few days away from the issue, I started looking at the creation of the EBS device script again with fresh eyes and noticed I was omitting the snapshot-id option. So it was indeed creating a blank EBS device. After adding the snapshot-id option, I'm seeing the partitions present as expected.

AWS_SamM · Answer

Hello, thank you for your post. I understand you have created a snapshot of a root volume of a RHEL 8 EC2 instance. You then created a new EBS volume from the snapshot and then attached the new EBS volume to a RHEL 8 EC2 instance as a secondary volume. After this, lsblk does not show any partitions on the new EBS volume, even though there were multiple partitions originally.

I was unable to reproduce this issue when following the same procedure.

I launched a new EC2 instance using a RHEL 8 AMI, then created a snapshot of its root volume. From the snapshot I created a new EBS volume. I then launched a new EC2 instance using a RHEL 8 AMI again, and attached the EBS volume from the previous step. When I ran the lsblk command, I was able to see that all of the original partitions were present and visible for the secondary volume:
```
# lsblk
NAME        MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
nvme0n1     259:0    0  10G  0 disk
├─nvme0n1p1 259:1    0   1M  0 part
└─nvme0n1p2 259:2    0  10G  0 part /
nvme1n1     259:3    0  10G  0 disk
├─nvme1n1p1 259:4    0   1M  0 part
└─nvme1n1p2 259:5    0  10G  0 part
```
I encourage you to open a support case so that one of our support engineers can inspect the details of the specific snapshot and EBS volume in question, and provide further assistance with this issue.

Why does attaching a non-root EBS drive created from a snapshot not maintain the partitions?

Relevant content