Hello.
Please help. How to bring the updated server (name PVE2) back to the cluster with all virtual machines on the server?
A cluster of 3 servers was assembled.
When updating packages of 1 server (name PVE2), the server does not connect to the cluster, it works as a separate server, there is access to it on port 8006, virtual machines are available on it.
The following messages are recorded in the cluster every second (name PVE1):
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Incoming packet has different crypto type. Rejecting
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Received message has invalid digest ... ignoring.
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Invalid packet data
Old packages, before the update:
corosync-2.4.1
libcorosync-2.4.1
pve-http-server-2.0.11
pve-container-2.0.24
pve-cluster-5.0.27
pve-qemu-server-5.0.27
pve-vncterm-1.5.2
pve-qemu-common-2.12.1
pve-access-control-5.0.8
pve-docs-5.2.3
pve-ha-manager-2.0.5
pve-qemu-system-2.12.1
pve-lxc-3.0.2
pve-qemu-aux-2.12.1
pve-spiceterm-3.0.4
pve-firewall-3.0.12
pve-guest-common-2.0.17
pve-novnc-0.6.4
pve-xtermjs-1.0.5
pve-storage-5.0.23
pve-qemu-img-2.12.1
pve-manager-5.2.3
pve-common-5.0.31
New packages after update (name PVE2):
libcorosync-3.0.4
corosync-3.0.4
pve-novnc-1.0.0
pve-qemu-server-6.2.2
pve-lxc-4.0.3
pve-guest-common-3.0.10
pve-qemu-system-5.1.0
pve-docs-6.2.4
pve-http-server-3.0.5
pve-cluster-6.1.8
pve-ha-manager-3.0.9
pve-storage-6.1.7
pve-manager-6.2.4
pve-xtermjs-3.10.1
pve-access-control-6.1.2
pve-container-3.1.5
pve-spiceterm-3.1.1
pve-qemu-img-5.1.0
pve-vncterm-1.6.2
pve-firewall-4.1.2
pve-qemu-common-5.1.0
pve-common-6.1.4
Config file from working corosync cluster:
cat /etc/pve/corosync.conf
logging {
debug: off
to_syslog: yes
}
nodelist {
node {
name: PVE1
nodeid: 1
quorum_votes: 1
ring0_addr: 192.168.10.181
}
node {
name: PVE2
nodeid: 2
quorum_votes: 1
ring0_addr: 192.168.10.182
}
node {
name: PVE3
nodeid: 3
quorum_votes: 1
ring0_addr: 192.168.10.183
}
}
quorum {
provider: corosync_votequorum
}
totem {
cluster_name: su-pve
config_version: 3
interface {
bindnetaddr: 192.168.10.181
ringnumber: 0
}
ip_version: ipv4
secauth: on
version: 2
}
For testing, we configured one more server with updated packages, to try to add to the existing cluster, when running the command:
pvecm add 192.168.10.181 --ring0_addr 192.168.10.184
the message comes out:
Unknown option: ring0_addr
400 unable to parse option
It turns out that instead of "--ring0_addr" you need to use "--link0"?
pvecm add 192.168.10.181 -link0 192.168.10.184
Maybe the problem is why the updated server (name PVE2) is not connecting back to the cluster?
Sorry for my translation, I use "https://translate.google.ru/".
Please help. How to bring the updated server (name PVE2) back to the cluster with all virtual machines on the server?
A cluster of 3 servers was assembled.
When updating packages of 1 server (name PVE2), the server does not connect to the cluster, it works as a separate server, there is access to it on port 8006, virtual machines are available on it.
The following messages are recorded in the cluster every second (name PVE1):
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Incoming packet has different crypto type. Rejecting
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Received message has invalid digest ... ignoring.
Apr 23 17:02:22 PVE1.localdomain corosync [1587]: [TOTEM] Invalid packet data
Old packages, before the update:
corosync-2.4.1
libcorosync-2.4.1
pve-http-server-2.0.11
pve-container-2.0.24
pve-cluster-5.0.27
pve-qemu-server-5.0.27
pve-vncterm-1.5.2
pve-qemu-common-2.12.1
pve-access-control-5.0.8
pve-docs-5.2.3
pve-ha-manager-2.0.5
pve-qemu-system-2.12.1
pve-lxc-3.0.2
pve-qemu-aux-2.12.1
pve-spiceterm-3.0.4
pve-firewall-3.0.12
pve-guest-common-2.0.17
pve-novnc-0.6.4
pve-xtermjs-1.0.5
pve-storage-5.0.23
pve-qemu-img-2.12.1
pve-manager-5.2.3
pve-common-5.0.31
New packages after update (name PVE2):
libcorosync-3.0.4
corosync-3.0.4
pve-novnc-1.0.0
pve-qemu-server-6.2.2
pve-lxc-4.0.3
pve-guest-common-3.0.10
pve-qemu-system-5.1.0
pve-docs-6.2.4
pve-http-server-3.0.5
pve-cluster-6.1.8
pve-ha-manager-3.0.9
pve-storage-6.1.7
pve-manager-6.2.4
pve-xtermjs-3.10.1
pve-access-control-6.1.2
pve-container-3.1.5
pve-spiceterm-3.1.1
pve-qemu-img-5.1.0
pve-vncterm-1.6.2
pve-firewall-4.1.2
pve-qemu-common-5.1.0
pve-common-6.1.4
Config file from working corosync cluster:
cat /etc/pve/corosync.conf
logging {
debug: off
to_syslog: yes
}
nodelist {
node {
name: PVE1
nodeid: 1
quorum_votes: 1
ring0_addr: 192.168.10.181
}
node {
name: PVE2
nodeid: 2
quorum_votes: 1
ring0_addr: 192.168.10.182
}
node {
name: PVE3
nodeid: 3
quorum_votes: 1
ring0_addr: 192.168.10.183
}
}
quorum {
provider: corosync_votequorum
}
totem {
cluster_name: su-pve
config_version: 3
interface {
bindnetaddr: 192.168.10.181
ringnumber: 0
}
ip_version: ipv4
secauth: on
version: 2
}
For testing, we configured one more server with updated packages, to try to add to the existing cluster, when running the command:
pvecm add 192.168.10.181 --ring0_addr 192.168.10.184
the message comes out:
Unknown option: ring0_addr
400 unable to parse option
It turns out that instead of "--ring0_addr" you need to use "--link0"?
pvecm add 192.168.10.181 -link0 192.168.10.184
Maybe the problem is why the updated server (name PVE2) is not connecting back to the cluster?
Sorry for my translation, I use "https://translate.google.ru/".
