automatic failover VM

bob360

New Member
Dec 17, 2015
3
1
3
33
Hello,

I am running proxmox 4.1 with 3 nodes in my cluster. I can migrate manually the VM from one host to another.

Is it possible to configure proxmox in order to automatically migrate a VM from node 1 to node 2 ? If my node 1 crashes because of the hardware for example.

Thanks,

Bob
 
Thank you for your answer, in the documentation I see this :

"If you need to reboot a node, e.g. because of a kernel update, you need to migrate all VM/CT to another node or disable them."

What happens if your server crashes ? You will not be able to disable or migrate them if the host is down ?

Maybe I missed something.

Bob
 
  • Like
Reactions: amil
Thank you for your answer, in the documentation I see this :

"If you need to reboot a node, e.g. because of a kernel update, you need to migrate all VM/CT to another node or disable them."

What happens if your server crashes ? You will not be able to disable or migrate them if the host is down ?

Maybe I missed something.

Bob

I completely agree with this question. For a non expected fatal node crash what should we expect or how should we act?
 
"If you need to reboot a node, e.g. because of a kernel update, you need to migrate all VM/CT to another node or disable them."

Please permit me another doubt regarding this afirmation: i suppose if we have a cluster with 3+ nodes for HA with proxmox, COULD tolerate 1 server node down, so about real HA.

1. If we have to schedule a node manteinance and migrate all vms from affected container to another or disable containers, where is the HA?

2. In case of manteinance of all nodes on cluster, we need to do the same on every node (migrate vms or stop disable). Depending on number of nodes this will be a headhache and the same.

3. In the case previously i comment: if a hardware node (in the cluster) it has a fatal crash (vms was not being migrated) what we should expect then. Pray and look for a recent backup? :rolleyes::(:(

Greetings!
 
I think it is possible to restart automatically VMs to another nodes in case of failure. Here is a demonstration under proxmox 3.2

Bob
 
1. If we have to schedule a node manteinance and migrate all vms from affected container to another or disable containers, where is the HA?

You have to define, what real HA means to you:
Real HA for me running a machine on two nodes with identical RAM content such that the machine will always work consistent if only one node fails. HA in proxmox means, that the machine will get started again automatically if the node crashes. You have to migrate the machines (online) manually.


2. In case of manteinance of all nodes on cluster, we need to do the same on every node (migrate vms or stop disable). Depending on number of nodes this will be a headhache and the same.

Why? Automate via commandline via
qm migrate

I often do node upgrades and migrate everything out of the way. Works like a charm and is a simple one-liner:

3. In the case previously i comment: if a hardware node (in the cluster) it has a fatal crash (vms was not being migrated) what we should expect then. Pray and look for a recent backup? :rolleyes::(:(

VMs are not migrated online, but they start automatically if the VM is marked as HA VM.
 
  • Like
Reactions: amil
[...]
1. If we have to schedule a node manteinance and migrate all vms from affected container to another or disable containers, where is the HA?

2. In case of manteinance of all nodes on cluster, we need to do the same on every node (migrate vms or stop disable). Depending on number of nodes this will be a headhache and the same.
[...]

afaik they decided to change that. check this post from Dietmar from this Morning (that dude is seriously early at work .. or late .. .)
https://forum.proxmox.com/threads/proxmox4-ha-not-working-feedback.23770/#post-126158
 

You have to define, what real HA means to you:
Real HA for me running a machine on two nodes with identical RAM content such that the machine will always work consistent if only one node fails. HA in proxmox means, that the machine will get started again automatically if the node crashes. You have to migrate the machines (online) manually.

Yes, just i refer. If one node fail (a unexpected down/crash or manteinance) all VMs can run consistent.

Why? Automate via commandline via
qm migrate

I often do node upgrades and migrate everything out of the way. Works like a charm and is a simple one-liner:
Yes, it´s very nice in case of manteinance,i s not the concept of ie. "vzmigrate"

You give me a bunch light on this sense, please correct me if i am wrong or i did not understand the concept:

In case one hardware node fail/crash, due identical RAM content, VMs is NOT being affected with a downtime (or downtime is minimal) due on RAM VMs are "live".
So in case of disaster on one node, we could ie. build a new node on new hardware to add to cluster and make a migration via proxmox to a "real storage" (qm migrate again)?


Greetings!
 
Yes, just i refer. If one node fail (a unexpected down/crash or manteinance) all VMs can run consistent.


Yes, it´s very nice in case of manteinance,i s not the concept of ie. "vzmigrate"

You give me a bunch light on this sense, please correct me if i am wrong or i did not understand the concept:

In case one hardware node fail/crash, due identical RAM content, VMs is NOT being affected with a downtime (or downtime is minimal) due on RAM VMs are "live".
So in case of disaster on one node, we could ie. build a new node on new hardware to add to cluster and make a migration via proxmox to a "real storage" (qm migrate again)?


Greetings!

Yet this identical RAM content HA is IMHO not available in KVM. If one node crashes and had a VM on it which was set as HA VM, the VM is automatically restarted on another node. This implies that you have a shared storage, otherwise you'll not have the VM data on the other node.
 
Yet this identical RAM content HA is IMHO not available in KVM. If one node crashes and had a VM on it which was set as HA VM, the VM is automatically restarted on another node. This implies that you have a shared storage, otherwise you'll not have the VM data on the other node.
qemu team is working on it, I think it should be ready for qemu 2.6 or 2.7
http://wiki.qemu.org/Features/COLO
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!