Add snapshot=0 option to Hard Disk ?

[Belokan]

Hello,

Could it be possible to implement a flag like the existing backup=0/1 and replication=0/1 but for snapshot ?

I have a OpenMediaVault VM with its boot disk based on CEPH and with 4 disks passed through.
I'd like to snapshot the boot disk before OS maintenance/upgrade/tests/whatever and the option is greyed out.

As I've backup & replication disabled the 4 "physical" disks why shouldn't I be able to snap the boot disk ? Maybe you should add another tick box for "snapshot" true/false in hard disk option or let snapshot to be taken on the "snapshot compliant" devices only ...

Here's my VM configuration:

#OpenMediaVault 3
agent: 1
bootdisk: virtio0
cores: 2
cpu: host
ide2: none,media=cdrom
memory: 4096
name: vmomv1
net0: virtio=7E:6F:E9:E8:3B:C0,bridge=vmbr0
net1: virtio=B6:F1:4E:C1:A7:61,bridge=vmbr1
numa: 0
onboot: 1
ostype: l26
protection: 1
scsihw: virtio-scsi-pci
smbios1: uuid=4f89b895-e7a0-46ee-a95f-6a441a116191
sockets: 1
virtio0: ceph-pve:vm-114-disk-1,size=16G
virtio1: /dev/disk/by-id/ata-WDC_WD20EARS-00MVWB0_WD-WMAZA2958354,backup=0,replicate=0,size=1953514584K
virtio2: /dev/disk/by-id/ata-SAMSUNG_HD204UI_S2H7JD2B106099,backup=0,replicate=0,size=1953514584K
virtio3: /dev/disk/by-id/ata-WDC_WD20EZRX-00D8PB0_WD-WMC4M2121278,backup=0,replicate=0,size=1953514584K
virtio4: /dev/disk/by-id/ata-SAMSUNG_HD204UI_S2H7JD2B185299,backup=0,replicate=0,size=1953514584K​

Thanks in advance !

 

[fabian]

https://bugzilla.proxmox.com/show_bug.cgi?id=1007

 

[Belokan]

Hi Fabian,

I've seen this post already but didn't found it relevant in the first place.
Does this mean that CT+bind = VM+pass-through regarding by question ?

So if I comment out virtio1 -> 4 lines and then try to snapshot again it will work (only on virtio0) ? When in the process do I have to uncomment the lines ? Right after the snapshot is created ? And when I'll have to rollback or remove the snapshot, will I have to comment out the 4 lines again ?

Thanks and cheers.

 

[fabian]

Hi Fabian,

I've seen this post already but didn't found it relevant in the first place.
Does this mean that CT+bind = VM+pass-through regarding by question ?

So if I comment out virtio1 -> 4 lines and then try to snapshot again it will work (only on virtio0) ? When in the process do I have to uncomment the lines ? Right after the snapshot is created ? And when I'll have to rollback or remove the snapshot, will I have to comment out the 4 lines again ?

Thanks and cheers.

it is on our TODO list to implement a "skip disk/mountpoint for snapshot purposes" option for containers and VMs, but it hasn't been done yet. that bug is the tracking bug.

 

[Kevo]

I also could benefit from an ignore disk in the VM config so I could take a snapshot of a boot volume. I currently also exclude my passthrough disk from backup as it is itself a backup disk inside the VM. Luckily my boot volume is small and it's not that big of a deal to back it up before trying out some changes, but being able to make snapshots work would be nice.

 

[Hartwin]

it is on our TODO list to implement a "skip disk/mountpoint for snapshot purposes" option for containers and VMs, but it hasn't been done yet. that bug is the tracking bug.

May it be possible to somehow slightly level up the priority of this task at the TODO? I'm facing the same problem: Created a NFS Storage at the host to use it as VitIO SCSI inside a VM (differential backups of windows servers needs dedicated volumes - impossible with NFS). These NFS Storage prevents me from creating snapshots because it's impossible to exclude it.

The mentioned bug at bugzilla is from 2016, your posting is from mid 2018 ... now the year is 2020 and there are some at least one more feature request at bugzilla.proxmox about exactly this feature:
https://bugzilla.proxmox.com/show_bug.cgi?id=2095

Is there any place to vote for features or the like?

Thanks
Hartwin

 

[alyarb]

I am replying just to say this is a much needed feature. live detach/attach is not realistic.

 

[fowr0yl]

very sad...
The bugzilla entry is about 6 years old. But no progress.

And in addition until now it's not possible to do snapshots even on ZFS datasets, if you just import a zPool ....