Proxmox

[hunter258]

Hi I am a complete newbie to proxmox.I've installed three nodes with HA and ceph. Live migration works perfectly.My only concern is when a node loses power or get shut down on the physical machine it will start up on another node but with a new startup again.I would just like for it to carry on without starting up again.How can i accomplish this if possible.Thanks Hunter

[Bob123]

It's been a while since playing with Proxmox but if it's anything like ESXi and Hyper-V then for HA to work that VM has to be shared between both machines.  Are you using one data store between the two nodes to store the VMs or are the VMs physically running on a machine.  I know when I did my POCs that I had a common data store that held the physical VMs and my two virtual nodes just pointed to them.  If they are physically on a machine then for HA to work you have to have a duplicate on a second node and maybe even a second network just for HA.  (Again remembering more of ESXi and Hyper-V).  Let me know if any of this helps or at least points you in the right direction.  If not I can try to setup a configuration likes yours and give it a try. 

[hunter258]

Hi

Are you using one data store between the two nodes to store the VMs - I am using a nfs server.

I do run a dedicated seperate network for this as well.

With all this in place when one node fails the second node does not take over but rather start from scratch again

as if it was offline.

Thanks

Hunty

[Bob123]

Hey Hunty,

Yes I only used one data store when doing my tests.  I was just wondering because I think in Hyper-V 2012's hypervisor they now have a feature that's called something like share nothing live migration which will actually take a running VM on a physical server and move it to another physical server while the VM is live.  And I thought for HA that had a feature like that as well where a backup or cloned VM sits on a second server just syncing the whole time until a failover occurs.  I'll see if I can setup a small proxmox lab to see if I can mimic this.  Can't promise how quickly I can do this since I just lost out on some really cheap hardware.  I think I have enough laying around to try this though.

Oh and give me as much info as possible.  What your using version wise.  Do I need to learn ceph as I've never used that before or can my data store just be an iSCSI or something like that?

Edited October 18, 2017 by Bob123
forgot to add a few remarks

[hunter258]

Hi

I am making use of proxmox 4.4-1 virtual environment.The latest version 5 is not as stable this is why i am running version 4.With regards to ceph no need to make use of it.

You can make use of nfs storage.On my research what I have done it sounds like I need to look for a fault tolerance system which i just cannot find anywhere.

If you can do live migration from one node to another node surely you should be able to run this seamlessly if one node suddenly shuts down without restarting.

Just shout if there is anything else you need and thanks for your advice thus far :)

Hunty

[Bob123]

So this is a bit weird and I'd like to get more information on this but according to proxmox's wiki: "Proxmox VE High Availability Cluster (Proxmox VE HA Cluster) enables the definition of high available virtual machines. In simple words, if a virtual machine (VM) is configured as HA and the physical host fails, the VM is automatically restarted on one of the remaining Proxmox VE Cluster nodes."

and:

You must meet the following requirements before you start with HA:

• at least three cluster nodes (to get reliable quorum)

 

I've never heard of needed 3 nodes not including a separate data store just for HA. 

I'll keep digging.

[hunter258]

Hi Bob

Once again thanks for your time and effort.Highly appreciated :)