Clustering of virtual machine of one proxmox

[Neelu Patel]

Hi,

I am new to this clustering concept, I have one proxmox with four VM(virtual machines) in it. I have to cluster these four VMs in such a way that if one VM fails another VM must start automatically.

Is it possible to do this task using the clustering concept and if yes then how can we do it.
can you please suggest a solution for this?

 

[sterzy]

Hi,

I think what you are looking for is our High Availability (HA) feature [1]. The basic idea is that you have multiple cluster nodes running PVE that share the storage for a given VM or Container. If one node fails, that will be detected, the node will separate itself from the cluster (self-fencing) and the VM will be migrated to another node in the cluster. This way the VM stays up. Although, do note that it might take a few minutes for node failure to be detect and migration to finish.

[1]: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#chapter_ha_manager

 

[Neelu Patel]

Yes, actually I know that we can make a cluster of several proxmox as a node with a combination of VMs into it. if any one proxmox or node fails another proxmox or node will start automatically.

but , I want to know whether HA is possible only on one proxmox where several vms are available and if one VM fails, will another VM in the same proxmox starts automatically or not.

 

[sterzy]

Hi,

So essentially you want to automatically reboot the VM in case it crashes? This is a separate issue from HA. To do this you need to set up a watchdog. A watchdog is basically a timer that when it expires carries out an action. So the idea here is to setup a watchdog that will, when it expires, restart the VM.

You will then need to configure the new hardware within the VM and also setup a daemon within that periodically resets the watchdog itself. How to do this depends on the OS you are running within the VM.

You can add a watchdog to an existing VM with the following command (replace <VM-ID> with the ID of the VM that you want to add a watchdog to):

qm set <VM-ID> --watchdog model=i6300esb,action=reset

This will add an emulate i6300esb watchdog to your machine that will reset the VM if the watchdog expires. After setting up the in the VM watchdog you will need to power off and restart your VM.

There are several more option you can configure (e.g. instead of resetting the VM you can just turn it off if it freezes). You can find more information about this in the manual [1].

[1]: https://pve.proxmox.com/pve-docs/pve-admin-guide.html#qm_options (scroll down to the last option "watchdog")

 

[Neelu Patel]

Ok, thanks for the help.