pmxcfs crit: can't initialize service

[bond347]

Hi All,

I have 3 hosts and planned to create a cluster.
Upon Cluster creation on my 1st node, i saw the following errors. FYI, i had rebooted the server too. Why I'm seeing these error messages?

Note: I have yet to add hosts 2 and 3 in this cluster.


# systemctl status pve-cluster.service
● pve-cluster.service - The Proxmox VE cluster filesystem
Loaded: loaded (/lib/systemd/system/pve-cluster.service; enabled; vendor preset: enabled)
Active: active (running) since Tue 2023-05-02 15:03:08 +08; 56min ago
Process: 1236 ExecStart=/usr/bin/pmxcfs (code=exited, status=0/SUCCESS)
Main PID: 1247 (pmxcfs)
Tasks: 6 (limit: 153953)
Memory: 66.8M
CPU: 1.445s
CGroup: /system.slice/pve-cluster.service
└─1247 /usr/bin/pmxcfs

May 02 15:58:13 pmxcfs[1247]: [dcdb] crit: can't initialize service
May 02 15:58:13 pmxcfs[1247]: [status] notice: start cluster connection
May 02 15:58:13 pmxcfs[1247]: [status] crit: cpg_initialize failed: 2
May 02 15:58:13 pmxcfs[1247]: [status] crit: can't initialize service
May 02 15:58:19 pmxcfs[1247]: [status] notice: update cluster info (cluster name cluster1, version = 1)
May 02 15:58:19 pmxcfs[1247]: [status] notice: node has quorum
May 02 15:58:19 pmxcfs[1247]: [dcdb] notice: members: 1/1247
May 02 15:58:19 pmxcfs[1247]: [dcdb] notice: all data is up to date
May 02 15:58:19 pmxcfs[1247]: [status] notice: members: 1/1247
May 02 15:58:19 pmxcfs[1247]: [status] notice: all data is up to date



EXTRA INFO:
===============================

# systemctl status corosync.service
● corosync.service - Corosync Cluster Engine
Loaded: loaded (/lib/systemd/system/corosync.service; enabled; vendor preset: enabled)
Active: active (running) since Tue 2023-05-02 15:58:14 +08; 6min ago
Docs: man:corosync
man:corosync.conf
man:corosync_overview
Main PID: 8877 (corosync)
Tasks: 9 (limit: 153953)
Memory: 127.6M
CPU: 2.218s
CGroup: /system.slice/corosync.service
└─8877 /usr/sbin/corosync -f

May 02 15:58:14 corosync[8877]: [QB ] server name: quorum
May 02 15:58:14 corosync[8877]: [TOTEM ] Configuring link 0
May 02 15:58:14 corosync[8877]: [TOTEM ] Configured link number 0: local addr: 10.110.14.11, port=5405
May 02 15:58:14 corosync[8877]: [KNET ] link: Resetting MTU for link 0 because host 1 joined
May 02 15:58:14 corosync[8877]: [QUORUM] Sync members[1]: 1
May 02 15:58:14 corosync[8877]: [QUORUM] Sync joined[1]: 1
May 02 15:58:14 corosync[8877]: [TOTEM ] A new membership (1.f) was formed. Members joined: 1
May 02 15:58:14 corosync[8877]: [QUORUM] Members[1]: 1
May 02 15:58:14 corosync[8877]: [MAIN ] Completed service synchronization, ready to provide service.
May 02 15:58:14 systemd[1]: Started Corosync Cluster Engine.



root@:~# pvecm status
Cluster information
-------------------
Name: cluster1
Config Version: 1
Transport: knet
Secure auth: on

Quorum information
------------------
Date: Tue May 2 16:09:29 2023
Quorum provider: corosync_votequorum
Nodes: 1
Node ID: 0x00000001
Ring ID: 1.f
Quorate: Yes

Votequorum information
----------------------
Expected votes: 1
Highest expected: 1
Total votes: 1
Quorum: 1
Flags: Quorate

Membership information
----------------------
Nodeid Votes Name
0x00000001 1 10.110.14.11 (local)





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

nodelist {
node {
name:
nodeid: 1
quorum_votes: 1
ring0_addr: 10.110.14.11
}
}

quorum {
provider: corosync_votequorum
}

totem {
cluster_name:
config_version: 1
interface {
linknumber: 0
}
ip_version: ipv4-6
link_mode: passive
secauth: on
version: 2
}

 

[fabian]

that's because pmxcfs starts before corosync, it does re-initialize the CPG stuff automatically once Corosync becomes available.

 

[bond347]

Hi Fabian,

Thanks for the fast response.
After leaving the server for a day, now, i don't see any errors.

Before creating a cluster, i noticed the pve-cluster was auto-started and the corosync.service was down. BTW, how do you start the pmxcfs? Was it pve-cluster.service?

 

[fabian]

yes, the "pve-cluster" service starts pmxcfs. corosync is only started if a corosync config exists, and that is only the case after you created the cluster (or joined an existing one of course).