New 3.1 user - stupid simple questions

C

candrews

New Member
Feb 28, 2014
2
0
1
I am very new to ProxMox. I have just installed 3.1 and set up to WinXP3 VMs. My question is about long term administration.
I understand that I can bring another proxmox server on as a cluster, and can move VMs to it. I am looking long term. My server has ProxMox installed on a single disk. IF that disk fails, I see disaster. With the cluster, are the VMs mirrored between the two nodes? As a failover? Or do I have to restore a backup onto the other node to bring a vm back online if main node fails?

ANy thoughts on this? Ideas on how you've ensured continuity?

Thanks!
 
candrews said:
I am very new to ProxMox. I have just installed 3.1 and set up to WinXP3 VMs. My question is about long term administration.I understand that I can bring another proxmox server on as a cluster, and can move VMs to it. I am looking long term. My server has ProxMox installed on a single disk. IF that disk fails, I see disaster. With the cluster, are the VMs mirrored between the two nodes? As a failover? Or do I have to restore a backup onto the other node to bring a vm back online if main node fails?ANy thoughts on this? Ideas on how you've ensured continuity?Thanks!
Click to expand...
You would need to restore a backup in your situation. If you want HA then you need some type of central storage or DRBD setup. This has been covered many many many times.
 
candrews said:
I am looking long term. My server has ProxMox installed on a single disk. IF that disk fails, I see disaster.
Click to expand...

You will. If you're really looking long term, and wish to avoid disaster, you *must* have affordable, server grade, redundant, protected hardware. And tested backups. Not jsut a single disk.
This is not about proxmox, obviously, is about "Long term" and "no disaster".

candrews said:
With the cluster, are the VMs mirrored between the two nodes?
Click to expand...

no, they're not. With proper working HA, you can have "failover" vms, in the sense that pve could automatically restart a failed vm on one node, on another node.

Marco
 
m.ardito said:
You will. If you're really looking long term, and wish to avoid disaster, you *must* have affordable, server grade, redundant, protected hardware. And tested backups. Not jsut a single disk.
This is not about proxmox, obviously, is about "Long term" and "no disaster".
Click to expand...

I had heard some bad news about having a hardware RAID, is that on machines that are being virtualized? Is there anything wrong with having a hardware raid on the ProxMox machine? Maybe a Raid1 set?
 
The basic idea of a real Proxmox cluster is to put all VMs on a separate Shared storage node such as CEPH, FreeNAS, OmniOS+Napp-It and use Proxmox nodes for just to run those VMs. This gives 100% redundancy for "Proxmox Cluster Nodes" only. If you want storage redundancy, obviously you will have to have Storage cluster. Hardware RAID or not, if you are going to put VMs on local storage where Proxmox is installed, you will not have seamless failover of VMs. Because in the event of a disaster, the entire VM will need to be copied to other node then start again. Whereas if the VM is on Shared Storage, its just matter of seconds for the VM to switch over and start again on good Proxmox nodes.
 
candrews said:
... and set up to WinXP3 VMs. ...
ANy thoughts on this? Ideas on how you've ensured continuity?
Click to expand...

And, adding to your search for "continuity", WinXP guests could not be your best choice... they're going to be totally unsupported starting next month, and there's a lot of fear that this OS will attract most threats as soon as connected to public networks, in any way.

Marco
 
You must log in or register to reply here.
Top Bottom