full system backup creating multiple disk

[fonefoo]

I'm running a full system backup on a VM. The VM itself is built with a single disk.
The backup runs fine, but when I go to veeam to export the disk, I find 2 disks...

1: Disk 0 sda 1 ( 20 GB )
2: Disk 1 swap: root ( 19.5 GB )

Is there anyway to consolidate these into a single disk, so that I can export them to be run in another hypervisor?
I've backed a few other machines, and haven't had this issue yet..


(from a centos 7 vm)

Code: Select all

# lsblk
NAME            MAJ:MIN RM  SIZE RO TYPE MOUNTPOINT
sda               8:0    0   20G  0 disk 
├─sda1            8:1    0  500M  0 part /boot
└─sda2            8:2    0 19.5G  0 part 
  ├─centos-root 253:0    0 17.5G  0 lvm  /
  └─centos-swap 253:1    0    2G  0 lvm  [SWAP]
sr0              11:0    1 56.3M  0 rom  

[nielsengelen]

Could you give some more info on how the backup job is configured ?

[fonefoo]

I used the veeam gui to setup the job, from within the vm. Backup mode "Entire machine". Destination VBR server.

Code: Select all

# veeamconfig job info --name VM_Backup
Backup job
   UUID: {3e37f8a8-3e3b-4029-a80f-a730b921c18a}
   Name: VM_Backup
   Repository UUID: {c4938cf1-3d92-4018-a375-ac2514198569}
   Repository name: [VEEAM]iSAN 3224 1TB Volume Linux
   Creation time: 1970-01-01 00:00:00
   Options:
      Compression: Lz4
      Deduplication: no
      Max Points: 14
   Objects for backup:
   Include All System: 

[PTide]

Have you tried to export any of them? Disk0 exports should give you the original 20 GB SDA image with all partitions intact, could you give it a try please?

[fonefoo]

Yeah I've exported them and tried them out. Both images when converted to raw are roughly 20GB. Disk 0 will boot to the kernel selection though fail shortly after. Disk 1 does not boot at all.

Here is what my agent log looks like..


The export


Export Session (exported to VHD format)


I used qemu-img to convert the vhds to a raw files, to be booted up into KVM. The machine gets me to the kernel sections, then fails shortly after.
This is the same process I've used on about 6 other VMs that has worked flawlessly. This particular VM is the only one that is create multiple disks, despite the fact that is only has one drive.

[PTide]

What's the version of VBR that you use? Is it v9.5 GA? If so then please keep in mind that GA does not support VAL until U1. Anyway please provide logs from /var/log/veeam and also please include dmesg -T output.

Thanks

[fonefoo]

If I boot disk0, this is what I get

[fonefoo]

What's the version of VBR that you use? Is it v9.5 GA? If so then please keep in mind that GA does not support VAL until U1. Anyway please provide logs from /var/log/veeam and also please include dmesg -T output.

Thanks

I'm running B&E 9.5.0.711

Here is my dmesg from the vm I'm backing up

Here is zip of /var/log/veeam from the vm I'm backing up

[PTide]

Please briefly describe all the machines that you backup with VAL. Are all of the machines identical or the failing one is anyhow special?

[fonefoo]

the other vm's I've backed up are ubuntu 14 and 16. they backed up and restored fine.

this will be the first centos vm I've tried to backup. I originally had difficulty installing veeamsnap, and ended up having to update my kernel before it'd install (kernel was older than the installed headers).
I'm attempting an install on another centos vm, though could eventually encounter something similar. It seems that when installing dkms newer headers (than the installed kernel) are installed as a dkms pre-req. then veeamsnap won't install because kernel and headers don't match.

[fonefoo]

quick update...
so I was starting to think that maybe my vm was the issue (I didn't build it, so I couldn't say for sure)

anyway I fired up a fresh installation of centos 7, installed the agent, ran a full system backup (no errors).
I'm still seeing (2) disks created, even though the vm only has a single disk.

took disk 0, converted the vhd to raw (using qemu-img). built another vm, attached the freshly converted raw. the vm begins to boot and just hangs just like the vm I started this thread with.
at the point the issue looks like it's specific to centos 7. I have not tried a recovery with 6 yet.

as mentioned, all my ubuntu vm's using the method above have restored perfectly (their backups only consist of a single disk, the centos vm's are creating 2).

at this point, I would invite any of you to please give it a shot and see if you're seeing the same thing.

[PTide]

Do Ubuntu machines also have LVM on them?

[fonefoo]

they have lvm2 installed, but their partitions were not configured for for lvm during the installation of the vm.

[PTide]

I've reporduced the issue. The information has been passed to development team.

Thank you.

[fonefoo]

great, thanks for helping me dig into this

[PTide]

UPDATE: It turns out there is no quick and easy way to fix that, so I can not give you any time estimates. Please try to use Recovery Media to restore the backup onto KVM virtual machine, and let me know how it works.

Thanks

[fonefoo]

ah you're right. GA didn't fix the issue I had, but I was able to successfully recover the same machine using the recovery ISO.