Cluster proxmox debian encrytp

[ocerda]

It has nothing to do with the disk being encrypted, I installed debian without encrypting and then proxmox and when creating the cluster and adding the second node exactly the same error.

Are you sure of a proxmox team that does not need any more packages?

 

[morph027]

I'm pretty sure they know as Proxmox is based of the same system, just a different installer

I can try to tinker with the problem tomorrow in our lab.

 

[ocerda]

I'm pretty sure they know as Proxmox is based of the same system, just a different installer

I can try to tinker with the problem tomorrow in our lab.

I would appreciate it, the situation is as follows:

1- I installed debian 9, when I get to the type of disk partitioning, I select lvm encrypted storage and everything else I leave by default.
When I get to packages I remove the graphic environment and I only leave minimal services of debian.

2- Once installed I follow this post to install proxmox.
https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_Stretch

And then the process to create the cluster, in this version you can create via web, but it does not matter if you create it via web or by commands, in both cases I get the same error.

Look to see if you can get it and thank you very much for your help.

 

[morph027]

Okay, tried and it worked. Installed with Debian with english language.

Installation failed first time with errors, but that was due to wrong /etc/hosts.

- Make sure to follow the steps: https://pve.proxmox.com/wiki/Instal...d_an_.2Fetc.2Fhosts_entry_for_your_IP_address
- DO NOT LEAVE THE "127.0.1.1 ...." entry in /etc/hosts
- Doublecheck if "hostname --ip-address" returns just one and the correct IP address

Different step I took:

- I adjusted the network config before installing proxmox (added vmbr0 to /etc/network/interfaces manually)

for the record, this is how my corosync file looks like:

logging {
  debug: off
  to_syslog: yes
}

nodelist {
  node {
    name: pve-1
    nodeid: 1
    quorum_votes: 1
    ring0_addr: 10.0.42.1
  }
}

quorum {
  provider: corosync_votequorum
}

totem {
  cluster_name: test
  config_version: 1
  interface {
    bindnetaddr: 10.0.42.1
    ringnumber: 0
  }
  ip_version: ipv4
  secauth: on
  version: 2
}

 

[ocerda]

The fault is when trying to add another machine to that cluster that you have created.

The file that you have passed, you only have one node, which will be where you created the cluster. That also worked for me, what does not work is to add a second node

 

[morph027]

Jip, this was the file before adding all the other nodes. I've added 2 other nodes and the cluster was working.

 

[ocerda]

Can you pass the file with the two nodes working to see how it looks?

 

[morph027]

Status:

Quorum information
------------------
Date:             Fri Sep  7 14:51:36 2018
Quorum provider:  corosync_votequorum
Nodes:            3
Node ID:          0x00000001
Ring ID:          1/12
Quorate:          Yes

Votequorum information
----------------------
Expected votes:   3
Highest expected: 3
Total votes:      3
Quorum:           2  
Flags:            Quorate

Membership information
----------------------
    Nodeid      Votes Name
0x00000001          1 192.168.1.1 (local)
0x00000002          1 192.168.1.2
0x00000003          1 192.168.1.3

Config:

logging {
  debug: off
  to_syslog: yes
}

nodelist {
  node {
    name: pve-1
    nodeid: 1
    quorum_votes: 1
    ring0_addr: 192.168.1.1
  }
  node {
    name: pve-2
    nodeid: 2
    quorum_votes: 1
    ring0_addr: 192.168.1.2
  }
  node {
    name: pve-3
    nodeid: 3
    quorum_votes: 1
    ring0_addr: 192.168.1.3
  }
}

quorum {
  provider: corosync_votequorum
}

totem {
  cluster_name: test
  config_version: 3
  interface {
    bindnetaddr: 192.168.1.1
    ringnumber: 0
  }
  ip_version: ipv4
  secauth: on
  version: 2
}

 

[ocerda]

I'll try it with the English language ... but come on, there's no point in failing for the language, it's the only different thing you have to me.

 

[morph027]

I don't think so, but hey, i've seen many bogus things in my career already

 

[morph027]

Okay, just tried with spanish debian, also works without hassle here.

Can you post your /etc/hosts and /etc/network/interfaces before setting up the cluster?

 

[ocerda]

This afternoon when I go to the office I send what you have asked me, at 15.30 Spanish time I publish it to you.

I will also send the configuration of the corosync.conf
The state of the nodes.

Thank you

 

[ocerda]

/etc/hosts

127.0.0.1 localhost
192.168.0.2 node1.infotiming.es node1 pvelocalhost
192.168.0.4 node3.infotiming.es node3 pvelocalhost

/etc/network/interfaces

source /etc/network/interfaces.d/*

auto lo
iface lo inet loopback

allow-hotplug enp1s0

iface enp1s0 inet dhcp

auto vmbr0
iface vmbr0 inet static
address 192.168.0.2
netmask 255.255.255.0
gateway 192.168.0.1
bridge-ports enp1s0
bridge-stp off
bridge-fd 0

can not pass the data from the corosync.conf or the state of the node, since the server has been taken by my partner today to do other tests.

I send you the host and network that they have

 

[morph027]

Looks good so far. Does

hostname --ip-address

return the correct ip of the server?

 

[ocerda]

Yes, that's the first thing I try and if it works, that's why I say I do not understand why it does not work.
It's not configuration failure, and I do not find out what it's

 

[Stoiko Ivanov]

the lines:

allow-hotplug enp1s0

iface enp1s0 inet dhcp

shouldn't be needed - do you have an additional ip configured on enp1s0?

 

[ocerda]

No, I know it can be manual instead of dhcp, but even that change does not work either.

 

[ocerda]

My problem has already been solved, I had in / etc / network / interfacez an interface of proxmox by dhcp and the other by manual, I thought that both were by dhcp and that has been the cause of always giving me the same error.

change the two and I put them manual.

Stoiko Ivanov reviewing what you said, I found out my mistake.

Thank you very much everyone, the clusters have already joined.

 

[Stoiko Ivanov]

Glad to hear!

 

[ocerda]

It is interesting these tests that I am doing, right now I have a cluster with encrypted disks, in case of theft they will not obtain any type of information.

My next step is to mount a 1TB encrypted disk to each tower and create a high availability shared ceph storage.

And if I get it, I will have 3 fully encrypted proxmox servers, sharing encrypted storage, and high availability when there are virtual machines that have crashes.