Hello,
We have a failed Proxmox 4.4->5.0 upgraded node that makes all cluster nodes (remaining 4) to fence themselves whenever it comes online. Seems I will have to reinstall the node for whatever reason.
I've just noticed something strange in corosync.conf while backing up the configuration:
cat /etc/pve/corosync.conf
logging {
debug: off
to_syslog: yes
}
nodelist {
node {
name: nthl12
nodeid: 1
quorum_votes: 1
ring0_addr: nthl12
}
node {
name: ntvl18
nodeid: 2
quorum_votes: 1
ring0_addr: ntvl18
}
node {
name: nthl11
nodeid: 4
quorum_votes: 1
ring0_addr: nthl11
}
node {
name: nthl03
nodeid: 3
quorum_votes: 1
ring0_addr: nthl03
}
node {
name: nthl16
nodeid: 5
quorum_votes: 1
ring0_addr: nthl16
}
}
quorum {
provider: corosync_votequorum
}
totem {
cluster_name: px
config_version: 7
ip_version: ipv4
secauth: on
version: 2
interface {
bindnetaddr: 172.20.10.7
ringnumber: 0
}
}
The Totem bindnetaddr node/IP was removed from the cluster some time ago. Should I remedy this and how?
Second question - can I "pvecm delnode" the failing node (the one that destroys quorum and fails SAN multipaths after upgrade whenever it's up) , remove auth key and reinstall/join cleanly under same name and IP?
Thanks!
We have a failed Proxmox 4.4->5.0 upgraded node that makes all cluster nodes (remaining 4) to fence themselves whenever it comes online. Seems I will have to reinstall the node for whatever reason.
I've just noticed something strange in corosync.conf while backing up the configuration:
cat /etc/pve/corosync.conf
logging {
debug: off
to_syslog: yes
}
nodelist {
node {
name: nthl12
nodeid: 1
quorum_votes: 1
ring0_addr: nthl12
}
node {
name: ntvl18
nodeid: 2
quorum_votes: 1
ring0_addr: ntvl18
}
node {
name: nthl11
nodeid: 4
quorum_votes: 1
ring0_addr: nthl11
}
node {
name: nthl03
nodeid: 3
quorum_votes: 1
ring0_addr: nthl03
}
node {
name: nthl16
nodeid: 5
quorum_votes: 1
ring0_addr: nthl16
}
}
quorum {
provider: corosync_votequorum
}
totem {
cluster_name: px
config_version: 7
ip_version: ipv4
secauth: on
version: 2
interface {
bindnetaddr: 172.20.10.7
ringnumber: 0
}
}
The Totem bindnetaddr node/IP was removed from the cluster some time ago. Should I remedy this and how?
Second question - can I "pvecm delnode" the failing node (the one that destroys quorum and fails SAN multipaths after upgrade whenever it's up) , remove auth key and reinstall/join cleanly under same name and IP?
Thanks!