backup made system flake out

[sdpve]

Hi,

I'm new to proxmox and has an issue with my first backup job.
I have proxmox 5.1. I have an OMV VM (100). I passed a hard disk to OMV. In OMV I shared folders on that hardisk over nfs. In proxmox I mounted the nfs share. I then setup a backup. I selected to backup the OMV VM 5 minutes in the future. When backup started, things got wierd with the GUI. The node and VMs had "?" on them. The backup said it was going on. I checked the status and it mentioned the hard disk I passed to OMV. I started wondering if it was trying to backup the disk I passed to OMV to the disk I passed to OMV (recursion). I wasn't sure so I tried to stop the backup on the GUI and that didn't work. I then tried to kill the processes running the backup. There was 3 of them, 1 for compression, and I forget the other two. When that didn't work I did kill -9. The processes were killed.
The GUI was still flakey so I tried login/logging out and it still showed question marks and it showed the backup as still going on. So then I tried "reboot now" on command line and I lost access to the server. I'll have to go reboot it manually.
Anyhow, my question is, when you backup a VM does it backup all the drives within it? Should I need backup a VM if I pass hardware to it? Is there a way on the GUI to schedule a backup of the VM without the drives in it or do I need to use command line?
Also, is there an option to backup the promox config itself?

Thanks,

 

[sdpve]

hmm, I've manually rebooted the server twice and can't ssh/ping to proxmox. This may have killed proxmox. Will have to hook up a monitor to it this weekend.

 

[Klaus Steinberger]

Btw. It is definitely not a good idea to mount a NFS Share which is running on a VM which is on the same host. What is first: Henn or Egg?

 

[sdpve]

Btw. It is definitely not a good idea to mount a NFS Share which is running on a VM which is on the same host. What is first: Henn or Egg?

how do you share a drive that you passed toi a VM then?

 

[Klaus Steinberger]

if you share back a VM's drive to the host running the VM you will always get into henn/egg trouble regardless of the protocol.

You can do it of course for short times for a transfer of data, but there are other ways to do that (scp or whatever). Backing up vm's into a VM running on the same host is a bad idea, think a little bit on it ....

 

[sdpve]

one thing that is unclear, if you backup a VM do you backup the configuration of the VM or do you backup the whole VM including hard disks in the VM? I'm just looking to backup the configuration.

I got a monitor on the server and it looks like the network connection is not working anymore. I didn't make any network changes so I'm not sure what happened with that. Is there a way to reset the network back to original state or should I just reinstall proxmox?

 

[Klaus Steinberger]

What we do:

we run backup weekly onto a external NFS Storage which in turn is backuped to a tape roboter (Tivoli agent, this is provided by the university computing center)

We backup state together with image, at least for alle images on our ceph storage.
some very large images (WSUS data etc.) sit on a external NFS Storage and are excluded from the vzdump backup. They are backuped by tivoli directly.

for your networking connection: Look into /etc/network/interfaces, it is just standard debian network configuration. should be easy to repair on command line.

 

[sdpve]

As far as backup, I figured out I can exclude the passed drive by setting "backup=0" on the VM config. That should solve my issue as I was able to make backups of the VM locally. But I havent tested it on the NFS share, since im still stuck on the networking issue (thinking of making a seperate thread for that). Anyhow, I think my backup issue is solved, so closing this thread if I can.