I'm trying to seperate the corosync network in our 3 node cluster (Proxmox 5.3-11) as described in https://pve.proxmox.com/wiki/Separate_Cluster_Network
After copying the new corosync.conf to /etc/corosync/corosync.conf I'm getting the following error:
According to https://pve.proxmox.com/wiki/Separate_Cluster_Network#quorum.expected_votes_must_be_configured thats because entries in the hosts-file(s) are wrong.
My hosts file on all 3 cluster nodes contains:
And corosync.conf looks like this:
All 3 names (coro0-proxmox-c1-n1, coro0-proxmox-c1-n2, coro0-proxmox-c1-n3) are resolvable on all 3 cluster nodes. I've already tried to use the corresponding IP addresses as ring0_addr's but I'm getting the same error.
Accroding to my research the problem could also be related to multicast problems but testing with omping doesn't show any problems:
Any ideas what I'm doing wrong?
After copying the new corosync.conf to /etc/corosync/corosync.conf I'm getting the following error:
Code:
Mar 19 10:22:57 proxmox-c1-n3 pmxcfs[3876]: [dcdb] notice: wrote new corosync config '/etc/corosync/corosync.conf' (version = 15)
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: notice [CFG ] Config reload requested by node 1
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: [CFG ] Config reload requested by node 1
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: crit [VOTEQ ] configuration error: nodelist or quorum.expected_votes must be configured!
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: crit [VOTEQ ] will continue with current runtime data
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: [VOTEQ ] configuration error: nodelist or quorum.expected_votes must be configured!
Mar 19 10:22:57 proxmox-c1-n3 corosync[3956]: [VOTEQ ] will continue with current runtime data
Mar 19 10:22:57 proxmox-c1-n3 pmxcfs[3876]: [status] notice: update cluster info (cluster name proxmox-c1, version = 15)According to https://pve.proxmox.com/wiki/Separate_Cluster_Network#quorum.expected_votes_must_be_configured thats because entries in the hosts-file(s) are wrong.
My hosts file on all 3 cluster nodes contains:
Code:
10.55.1.1 coro0-proxmox-c1-n1.mydomain.com coro0-proxmox-c1-n1
10.55.1.2 coro0-proxmox-c1-n2.mydomain.com coro0-proxmox-c1-n2
10.55.1.3 coro0-proxmox-c1-n3.mydomain.com coro0-proxmox-c1-n3And corosync.conf looks like this:
Code:
logging {
debug: off
to_syslog: yes
}
nodelist {
node {
name: proxmox-c1-n1
nodeid: 1
quorum_votes: 1
ring0_addr: coro0-proxmox-c1-n1
}
node {
name: proxmox-c1-n2
nodeid: 2
quorum_votes: 1
ring0_addr: coro0-proxmox-c1-n2
}
node {
name: proxmox-c1-n3
nodeid: 3
quorum_votes: 1
ring0_addr: coro0-proxmox-c1-n3
}
}
quorum {
provider: corosync_votequorum
}
totem {
cluster_name: proxmox-c1
config_version: 15
interface {
bindnetaddr: 10.55.1.0
ringnumber: 0
}
ip_version: ipv4
secauth: on
version: 2
}All 3 names (coro0-proxmox-c1-n1, coro0-proxmox-c1-n2, coro0-proxmox-c1-n3) are resolvable on all 3 cluster nodes. I've already tried to use the corresponding IP addresses as ring0_addr's but I'm getting the same error.
Accroding to my research the problem could also be related to multicast problems but testing with omping doesn't show any problems:
Code:
omping -c 10000 -i 0.001 -F -q coro0-proxmox-c1-n1 coro0-proxmox-c1-n2 coro0-proxmox-c1-n3
coro0-proxmox-c1-n2 : unicast, xmt/rcv/%loss = 10000/10000/0%, min/avg/max/std-dev = 0.047/0.127/0.283/0.036
coro0-proxmox-c1-n2 : multicast, xmt/rcv/%loss = 10000/10000/0%, min/avg/max/std-dev = 0.051/0.134/0.286/0.036
coro0-proxmox-c1-n3 : unicast, xmt/rcv/%loss = 10000/10000/0%, min/avg/max/std-dev = 0.049/0.143/0.304/0.042
coro0-proxmox-c1-n3 : multicast, xmt/rcv/%loss = 10000/10000/0%, min/avg/max/std-dev = 0.059/0.150/0.307/0.041Any ideas what I'm doing wrong?
