Veeam for Proxmox

[daschmidt94]

Good morning,

since the last month i have a Dell T30 with Proxmox running. There are 1 SSD for the system, 2 HDD with 2TB in ZFS mirror and 1 HDD 1TB for openmediavault.
Now I'm looking for a great backupsolution. With the research I found veeam. It looks nice and easy. Now my problem is how does it or whether it works at all.


best regards

[HannesK]

Hello,
and welcome to the forums.

I assume that you use "real" VMs and no paravirtualization. I would not try do backup anything on the hypervisor (proxmox) level in general. This limits restore options and it's hard get the virtual machines into a consistent state without shutting them down.

I recommend installing Veeam Agent for Windows or Linux (depending what you use) in the VMs.

Best regards,
Hannes

[daschmidt94]

thanks for your answere.

The most of the VM's are debian. Is there the best way, that i install the Veeam agent on the Vm's?

best regards

[HannesK]

Is there the best way, that i install the Veeam agent on the Vm's?

yes, that's what I wrote before.

[nielsengelen]

Depending on how your Proxmox is setup it may be that you can only use file-level backup though. If you are using QEMU, you can use all methods.

[daschmidt94]

thanks it works.

Last question is: i will backup my nextcloud is it possible to get in the maintance mode from the nextcloud for the backup or are there no problems without maintancemode

[nielsengelen]

I'm not familiar with how nextcloud works. If you could provide some more information on how or why maintenance mode is used, we may provide you an answer.

[daschmidt94]

The maintenance: mode prevents inconsistencies in your data, closes active sessions and prevents new logins.

[gideon pelser]

Hi Daschmidt94,

You should be able to add "sudo -u apache php /path/to/occ maintenance:mode --on" as a pre-schedule script / command and "--off" as a post schedule command in the agent. I think it is under advanced

[PTide]

@daschmidt94

Just keep in mind that there are actually two pairs of scripts - pre- and post-job, and pre-freeze/post-thaw.

The first one is executed before the backup session starts and after the backup session ends, the second one is available only in a server version and is executed right before snapshot and right after the snapshot is taken.

Thanks!

[Ejdesgaard]

The best workaround you can do within the Veeam ecosystem is to utilize the Veeam agent for Linux (I have been using it for a few years on non-vSphere/hyper-v solutions.)

If you need to backup Linux client orientated distributions such as Fedora, that are quite early adapters of new kernels, then you will get into trouble from time to time, because the kernel module is out-of-tree and trailing.
A really good example of this is a current pull-request, fixing support for the 5.6 kernel, it has been sitting around as a pull request for 20 days now with no visible action from Veeam at all.
https://github.com/veeam/veeamsnap/pull/5
This would be completely avoidable if Veeam decided to upstream the veeamsnap module (The license is already matching the kernel license)

A thread have been ongoing since 2014 asking about Proxmox (KVM) support for Veaam Backup & Replication
veeam-backup-replication-f2/is-veeam-do ... 43-60.html

If you are looking for hypervisor integration then you sadly need to look elsewhere, unless something has changed very recently VBR
@Gostev can you maybe elaborate on this for VBR ?

@Daschmidt94 a completely alternative route for your NC setup could be to leverage S3 compatable storage and use its builtin features directly for versioning/snapshotting etc. (CEPH comes to mind)

[PTide]

@Ejdesgaard

This would be completely avoidable if Veeam decided to upstream the veeamsnap module (The license is already matching the kernel license)

Well, we would be happy to do so. Unfortunately, it's not just a matter of us wanting do upstream : )

A really good example of this is a current pull-request, fixing support for the 5.6 kernel, it has been sitting around as a pull request for 20 days now with no visible action from Veeam at all.

Thank you for pointing that out. I'll ask out kernel guy to take a look

Thanks!

[Ejdesgaard]

@PTide
What do you mean with

Well, we would be happy to do so. Unfortunately, it's not just a matter of us wanting do upstream : )

If you upstreamed it, then the pull-request that's outstanding, would be trivial, because it would only be included in the releases moving forward unless you choose to backport to stable release, instead of now where you need to keep track of what kernel to use with what version of your codebase.
The supported distros and associated kernel support on https://helpcenter.veeam.com/docs/agent ... tml?ver=40 is quite a mess as a result of this.

Also, if you upsteamed veeamsnap then you would catch said pull request etc. during the kernel release-cycle instead of after a distribution release.

[PTide]

I mean that one does not simply upstream stuff. It can be a long road during which the code is thoroughly reviewed before it is accepted. In other words, we've already taken certain steps in order to upstream our module, it just won't happen as fast as we would like it to.

is quite a mess as a result of this.

Well, if you have any ideas on how to make it less messy then I am all ears : ) You are probably the first one to say that so far.

Thanks!

[furyflash]

We have production project based on VMs in Openstack + KVM public cloud. Currently 170 VMs managed by Veeam Backup and Replication and Enterprise Manager + Veeam Agents.
I have started personal video tutorial based on community edition for small customers.
It is valid for KVM, Proxmox, VirtualBox.
In some nearest video I will show how to restore Windows and Linux KVMs from from scratch if it isn't possible to mount ISO/CDROM device.

https://www.youtube.com/watch?v=XE5OCFd ... eGbYPfl8J5