Skip to main content

Mounting EBS SOP

Scenarios

  • EBS volume created from a snapshot.
  • Attaching an existing volume to another instance.
  • Attaching a new EBS volume to an instance.

Steps

Assuming that we already have an EBS volume to attach to an instance.

  • The device could be attached to the instance with a different device name than you specified while attaching the volume.
    • For example: /dev/sdf can be changed to /dev/nvme1n1, the new namind depends on how many existing volumes are currently attached.
  • Use lsblk -f to view available disks
    • Image
  • Checked whethere or not the new volume has a filesystem on it.
    • This can be checked either from the outout of the lsblk -f command in the FSTYPE column. As seen in this case, we have FSTYPE of xfs, which is correct since, this volume was create from a snapshot.
    • Or, file -s /dev/nvme1n1p1, here p1 denotes the partition. If you get an output where new volume isn't partitioned, then simple check filesystem for whole disk, file -s /dev/nvme1n1
      • Image
    • Ideally, a new volume shouldn't have a FileSystem on it.
    • While, if you're restoring from a snapshot, then it should as seen in above image. If it doesn't have a filesystem in this case, it means something when wrong with either the snapshot or while attaching the volume.
      • In this case, detach the volume from the AWS console.
      • Delete this and create a new volume from the snapshot and try again.
  • (Conditional) Create a file system with mkfs -t xfs /dev/nvme1n1
  • Create a mount point with the mkdir command, mkdir -p /mnt/newvol
  • Mount the new volume.
    • If mounting a volume which already had a partition make sure to mount the partition which had the filesystem.
      • mount /dev/nvmen1p1 /mnt/newvol
    • If mounting a new volume for which filesystem was create in above step.
      • mount /dev/nvmen1 /mnt/newvol
  • Make it permanent.

Common Error

  • wrong fs type, bad option, bad superblock on /dev/nvme1n1p1, missing codepage or helper program, or other error.

    • Check the mount logs with dmesg | tail

    • If you see, Filesystem has duplicate UUID errors. This means the volume you're mounting has the same ID as an existing volume. Follow the below additional setps to mount.

      • Error

      • Here

      • # mount without uuid
        mount -o nouuid /dev/nvme1n1p1 /mnt/newvol
        # now unmount
        umount /mnt/newvol
        # generate a new uuid
        xfs_admin -U generate /dev/nvme1n1p1
        #finally mount with uuid
        mount /dev/nvme1n1p1 /mnt/newvol

    • ref

Troubleshooting

This will help in while increasing the size of the ebs volume greater than 2TB.

  • If you are not able to increase the size of ebs volume greater than 2TB on server then check the current growpart package using below command.
    • Run this command rpm -qa| grep -i growpart and check output as shown below in ss.
      • Image
    • If the package version is 0.29 then we can't increase the size of ebs volume greater than 2TB.
    • In this case, Upgrade the growpart package version using below command. package version shoule be greater than or equal to 0.31.
      • sudo yum update cloud-utils-growpart
    • After upgrading the growpart package version then you can continue with the same steps to increase the volume size.