LXD drivers fail to start VMs if there are duplicated fs uuids in XFS images

[dann1]

Description
The LXD drivers need to mount the filesystem of the images ext* and XFS do not have their UUIDs regenerated when an instance is instantiated for the first time, causing a VM to be unable to start due to the mount action failing when the filesystem is XFS.

To Reproduce

• Deploy VM backed by a non-persistent XFS image
• Deploy it again and it will fail

The log will be similar to

Wed Mar 20 18:13:23 2019 [Z0][VM][I]: New state is ACTIVE
Wed Mar 20 18:13:23 2019 [Z0][VM][I]: New LCM state is PROLOG
Wed Mar 20 18:13:24 2019 [Z0][VM][I]: New LCM state is BOOT
Wed Mar 20 18:13:24 2019 [Z0][VMM][I]: Generating deployment file: /var/lib/one/vms/8/deployment.0
Wed Mar 20 18:13:24 2019 [Z0][VMM][I]: Successfully execute transfer manager driver operation: tm_context.
Wed Mar 20 18:13:24 2019 [Z0][VMM][I]: Successfully execute network driver operation: pre.
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: Command execution fail: cat << EOT | /var/tmp/one/vmm/lxd/deploy '/var/lib/one//datastores/0/8/deployment.0' 'ubuntu1604-context-lxd-9180d-0.test' 8 ubuntu1604-context-lxd-9180d-0.test
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: deploy: Using qcow2 mapper for /var/lib/one/datastores/0/8/disk.0
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: deploy: Mapping disk at /var/lib/lxd/storage-pools/default/containers/one-8/rootfs using device /dev/nbd1
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: deploy: Mounting /dev/nbd1 at /var/lib/lxd/storage-pools/default/containers/one-8/rootfs
Wed Mar 20 18:13:26 2019 [Z0][VMM][E]: deploy: mount_dev: mount: wrong fs type, bad option, bad superblock on /dev/nbd1,
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: missing codepage or helper program, or other error
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: 
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: In some cases useful info is found in syslog - try
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: dmesg | tail or so.
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: /var/tmp/one/vmm/lxd/deploy:72:in `<main>': failed to setup container storage (RuntimeError)
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: ExitCode: 1
Wed Mar 20 18:13:26 2019 [Z0][VMM][I]: Failed to execute virtualization driver operation: deploy.
Wed Mar 20 18:13:26 2019 [Z0][VMM][E]: Error deploying virtual machine
Wed Mar 20 18:13:26 2019 [Z0][VM][I]: New LCM state is BOOT_FAILURE

The core issue can be reproduced outside of the opennebula scope

• create a block device truncate -s 100M img.xfs
• format the block: mkfs.xfs img.xfs
• copy the block : cp img.xfs img_duplicated.xfs
• mount block1: mkdir /mnt/1 && mount img.xfs /mnt/1
• mount block1: mkdir /mnt/2 && mount img_duplicated.xfs /mnt/2

The latter will fail. Checking dmseg will output something similar to:

[182584.911629] XFS (loop13): Filesystem has duplicate UUID 9d6923a7-32ae-41c0-9c1e-e9d6064003e3 - can't mount

Expected behavior
Partitions won't fail to mount if they can be mounted

Details

• Affected Component: Virtualization Drivers
• Hypervisor: LXD
• Version: 5.8.0

Additional context
Mounting ext4 filesystems with duplicated uuids desn't fail.

Progress Status

• Branch created
• Code committed to development branch
• Testing - QA
• Documentation
• Release notes - resolved issues, compatibility, known issues
• Code committed to upstream release/hotfix branches
• Documentation committed to upstream release/hotfix branches