Node is online but appears offline - this only happens if a node in the cluster goes down

[byAd12]

Hello, I need help, please.

In my Proxmox cluster, if only one node goes down, all nodes appear offline except for the node I access through the web interface.
For example, if I access the management interface for the coruna1 node, I only see coruna1 as online, and the rest of the nodes are offline.
I can access the nodes that are marked as offline, but they are actually working.

In the Datacenter > Summary tab, it says that there is quorate: "Cluster: clusterVPN, Quorate: Yes"

/etc/pve/corosync.conf:

root@coruna1:~# /usr/bin/cat /etc/pve/corosync.conf
logging {
  debug: off
  to_syslog: yes
}

nodelist {
  node {
    name: coruna1
    nodeid: 1
    quorum_votes: 1
    ring0_addr: 172.16.0.100
  }
  node {
    name: malaga1
    nodeid: 3
    quorum_votes: 1
    ring0_addr: 172.16.0.102
  }
  node {
    name: malaga2
    nodeid: 2
    quorum_votes: 1
    ring0_addr: 172.16.0.103
  }
}

quorum {
  expected_votes: 1
  last_man_standing: 1
  provider: corosync_votequorum
  wait_for_all: 0
}

totem {
  cluster_name: clusterVPN
  config_version: 13
  interface {
    linknumber: 0
  }
  ip_version: ipv4-6
  link_mode: passive
  token: 15000
  secauth: on
  version: 2
}
root@coruna1:~#

In the images, the malaga2 node is offline and the rest are online.

 

[UdoB]

expected_votes: 1
last_man_standing: 1
wait_for_all: 0

Are you sure you know exactly what this does? I am sure, I do not ;-)

My recommendation: return to a default setup, then it will react with the default behavior.

 

[byAd12]

Are you sure you know exactly what this does? I am sure, I do not ;-)

My recommendation: return to a default setup, then it will react with the default behavior.

Hi, thanks for replying.

I've changed the configuration on the nodes to:

logging {
  debug: off
  to_syslog: yes
}

nodelist {
  node {
    name: coruna1
    nodeid: 1
    quorum_votes: 1
    ring0_addr: 172.16.0.100
  }
  node {
    name: malaga1
    nodeid: 3
    quorum_votes: 1
    ring0_addr: 172.16.0.102
  }
  node {
    name: malaga2
    nodeid: 2
    quorum_votes: 1
    ring0_addr: 172.16.0.103
  }
}

quorum {
  provider: corosync_votequorum
}

totem {
  version: 2
  cluster_name: clusterVPN
  secauth: on
  config_version: 15
  interface {
    linknumber: 0
  }
}

I also restarted the services on all nodes but it still appears as offline.

Attached is the output of the command "journalctl -u corosync -b".

I should also clarify that the nodes connect via Netbird VPN since they are in different locations, and that I am not using Proxmox HA. Therefore, it works remarkably well when all nodes are online, as this is the only time the entire node list turns green.

The "pvecm status" command gets stuck after the "votequorum information" section (this doesn't usually happen if all nodes are online). I also show this in another attached image.