One question about clustering in Proxmox 2.x

Hello, is it normal that I don't have write access to these files?

Code:
root@proxmox2:~# ll /etc/pve
total 5
-r--r----- 1 root www-data  451 Apr 20 14:08 authkey.pub
-r--r----- 1 root www-data  182 May 11 10:55 cluster.conf
-r--r----- 1 root www-data  236 May 11 10:55 cluster.conf.old
-r--r----- 1 root www-data   73 May  4 11:00 datacenter.cfg
lr-xr-x--- 1 root www-data    0 Jan  1  1970 local -> nodes/proxmox2
dr-xr-x--- 2 root www-data    0 Apr 20 14:08 nodes
lr-xr-x--- 1 root www-data    0 Jan  1  1970 openvz -> nodes/proxmox2/openvz
dr-x------ 2 root www-data    0 Apr 20 14:08 priv
-r--r----- 1 root www-data 1533 Apr 20 14:08 pve-root-ca.pem
-r--r----- 1 root www-data 1679 Apr 20 14:08 pve-www.key
lr-xr-x--- 1 root www-data    0 Jan  1  1970 qemu-server -> nodes/proxmox2/qemu-server
-r--r----- 1 root www-data  150 Apr 24 13:20 storage.cfg
-r--r----- 1 root www-data  181 May 11 11:09 user.cfg
-r--r----- 1 root www-data  119 Apr 26 09:36 vzdump.cron
 
Last edited:
if you loose quorum, /etc/pve is read-only.
 
thank you for your answer!

How can I verify if I loose quorum?
Is there any documentation I can read about quorum?
 
After reinstall proxmox in server 2 and reboot two nodes all its ok

Thanks a lot!
 
Mmm.. probably that documentation assumes one knows the concept of "quorum"... but this is not the case...
 
ok, I've read that a cluster quorum is the cluster configuration.. so, loosing a quorum means loosing the cluster configuration, right? But, where is the cluster configuration stored? If I loose quorum, how can I repair it?
 

A common method for defining quorum is to say that a subcluster of size n in a cluster of m total nodes has quorum if it has a plurality of nodes in its partition. That is, it has n members where n > INT(m/2). This simple method works quite well for larger clusters with generally reliable communications. However it breaks down for 2-node clusters, and may perform poorly for geographically dispersed clusters without highly reliable communications between sites.

so the problem could be that I have only 2 nodes?
 
Last edited:
I don't know.


I have a 2 cluster with proxmox 2.1. running.
I use udpu protocol for corosync instead multicast because my provider (OVH) not permite multicast (only with a extra cost).


After making written changes in this thread it works for me.
 
I'm inside a lan so I think a don't have multicat problems...

The problem is, I think, I have two nodes, let's say node1 and node2. Node1 was the main node and I tried to make node2 the main node... by adding/removing nodes, changing votes...

I hope that reinstalling isn't the only solution to solve my problem.
 
Last edited:
I'm inside a lan so I think a don't have multicat problems...

LAN does not mean that multicast is working. This depends on your switches, especially managed switches blocks multicast by default and needs to be enabled first.
 
Sorry for not responding. I found the solution to the problem 2 days ago and I forgot to post it here to share it.

After much research on google, and testing many things, I found a solution to my problem

Proxmox 2.x seems that generates a key for each node in the cluster.

Having to reinstall the system and reattach the new node with the same name that had earlier, this gives problems with the key that already has generated by master cluster.

Then I did the following:

In the master cluster i run these commands

# pvecm nodes (to view te name of the nodes to belong the cluster, and identify the name of the node that had problems)
# pvecm delnode name-of-the-node (to remove the node with problems of cluster)

I reinstall the proxmox on that node again, and I put the same node name that had originally.

In the new node i run these commands

# pvecm add IP-master-cluster

And it worked perfectly and no problems with the quorum.

Hope this helps others, Greetings and thanks for the aid.
 
I have configured a third host with Proxmox 2.1 named Node3. I have created the cluster on Node3 and added Node1 and Node2 to that cluster.
Now quorum is ok, /etc/pve is read-write and Node3 is the master.

Can I force Node1 to become master?
 
I have configured a third host with Proxmox 2.1 named Node3. I have created the cluster on Node3 and added Node1 and Node2 to that cluster.
Now quorum is ok, /etc/pve is read-write and Node3 is the master.

Can I force Node1 to become master?

there is no master, its a quorum based system.
 
each node has one vote by default. but you can change this is you know what you are doing.
 
You don't need to re-install the node. Save the current VM configs in /etc/pve/local/qemu-server before removing everything.
Remove the node from the cluster with pvecm delnode <node to be removed> on one of the other nodes.

Perform the following commands on the node to be removed:

service cman stop
killall -9 corosync cman dlm_controld fenced
service pve-cluster stop
rm /etc/cluster/cluster.conf
rm -rf /var/lib/pve-cluster/* /var/lib/pve-cluster/.*
rm /var/lib/cluster/*
reboot (sometimes cluster kernel modules remain hang on the connections so it is better to reboot, to be sure).


After rebooting, you can add the node as usual:
pvecm add <IP address of one of the nodes already in the cluster>


I've broken and re-did a cluster this way for about 7-8 times in the past two weeks (it is important to me to make sure that I can restore a broken cluster before we go live).
 
Last edited:

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!