Cluster with unicast

[Ch@rlus]

Hello,

As OVH doesn't support multicast, I decided to configure my cluster to use unicast.

So, I created a cluster on Node 1, changed conf to add unicast", and added the following line in /etc/hosts "IP.OF.NOD.2 nsXXXXXX.ovh.net nsXXXXXX" and rebooted the node.

I reinstalled node 2 (to be sure) and do "pvecm add IP.OF.NOD.1". As expected, I get "Waiting for quorum... Timed-out waiting for cluster". After that, I added thte following line in /etc/hosts ""IP.OF.NOD.1 nsXXXXXX.ovh.net nsXXXXXX" and rebooted the node.

But, after that, if I do "/etc/init.d/cman restart" on node 2, I get an error : Waiting for quorum... Timed-out waiting for cluster.

What i'm doing wrong ?

-Ch@rlus

 

[Ch@rlus]

Up ? :/

 

[snowman66]

http://forum.proxmox.com/threads/96...ead-of-multicast-(for-Proxmox-VE-2-x-Cluster)

 

[Ch@rlus]

Well,

That's what I did.

But now, new problem, after a fresh install of server 2, i can't do anything on it : Cman is not running and, when I try to launch it, It doesn't print anything :

root@server2:~# /etc/init.d/cman restart
root@server2:~#

Any idea ?

 

[tom]

works here is a test environment - we just slightly updated http://pve.proxmox.com/wiki/Multicast_notes#Use_unicast_instead_of_multicast
(added a hint for editing the cluster.conf).

 

[Ch@rlus]

Thanks for the reply but, now, i've got a new problem...

Juste maked a fresh install (on proxmox 2.3) and, cman cannot start at all.

When I try to do "apt-get install cman", it told me that there is problem with dependency :

Dépend: libccs3 (>= 3.0.12) mais ne sera pas installé
Dépend: libcorosync4 (>= 1.2.1)
Dépend: libfence4 (>= 3.0.12) mais ne sera pas installé
Dépend: libopenais3 (>= 1.1.2)
Dépend: corosync (>= 1.2.1)
Dépend: openais (>= 1.1.1)

When I try to do /etc/init.d/cman start, he doesn't return anything..

Any idea ?

 

[tom]

you install from ISO? no need to install cman with apt-get.

after installation, just start with cluster creation - 'pvecm create yourclustername'

see http://pve.proxmox.com/wiki/Proxmox_VE_2.0_Cluster

 

[Ch@rlus]

Well, cman haven't to be started when node is just reinstalled.

New problem now : I can't do "pvecm add XX", i've got : unable to copy ssh ID

Any idea ?

 

[Ch@rlus]

Well, I think I've found the problem.

Server1's cluster seems to be broked. Is there a way to recreate it ? (Wihout reinstall it :/)

 

[tom]

if you know what your are doing, yes.

e.g. http://pve.proxmox.com/wiki/Proxmox_VE_2.0_Cluster

 

[Ch@rlus]

Hi,

So now, I haven't got the error "Unable to copy Ssh id" but, still have "Waiting for quorum... Timed-out waiting for cluster".

/etc/host is the same on both nodes, with IP of all nodes into it.

cluster.conf is the ssame too :

<?xml version="1.0"?>
<cluster name="VPS" config_version="9">


<cman keyfile="/var/lib/pve-cluster/corosync.authkey" transport="udpu"/>


<clusternodes>
<clusternode name="ns390755" votes="1" nodeid="1"/>
<clusternode name="ns263964" votes="1" nodeid="2"/></clusternodes>


</cluster>

So, what is wrong ?!

EDIT : corosync.log => http://pastebin.com/8wsmmpwP

 

[Bestbeast]

Hi,

So now, I haven't got the error "Unable to copy Ssh id" but, still have "Waiting for quorum... Timed-out waiting for cluster".

/etc/host is the same on both nodes, with IP of all nodes into it.

cluster.conf is the ssame too :

<?xml version="1.0"?>
<cluster name="VPS" config_version="9">


<cman keyfile="/var/lib/pve-cluster/corosync.authkey" transport="udpu"/>


<clusternodes>
<clusternode name="ns390755" votes="1" nodeid="1"/>
<clusternode name="ns263964" votes="1" nodeid="2"/></clusternodes>


</cluster>

So, what is wrong ?!

EDIT : corosync.log => http://pastebin.com/8wsmmpwP

Any solution for this, I am running onto the same problem

 

[Bintelcom]

[Quote = "Bestbeast, mensaje: 156.745, miembro de: 38722"] Cualquier solución para esto, estoy corriendo en el mismo problema [/quote]
tambien problema, hay alguna solucion?