Proxmox Server Uptime Issues and Scheduled Reboots

[Zayy]

For the past few weeks, I’ve been experiencing poor uptime on my Proxmox servers. As a newcomer to Proxmox, I’m unsure of the exact cause, but the servers frequently go offline, requiring manual intervention to bring them back online. I’ve collected some system logs, and I’m open to trying various solutions to resolve this issue.
System Logs:

root@pve1:~# journalctl -p err -f
May 25 14:26:34 pve1 pmxcfs[771]: [quorum] crit: quorum_initialize failed: 2
May 25 14:26:34 pve1 pmxcfs[771]: [quorum] crit: can't initialize service
May 25 14:26:34 pve1 pmxcfs[771]: [confdb] crit: cmap_initialize failed: 2
May 25 14:26:34 pve1 pmxcfs[771]: [confdb] crit: can't initialize service
May 25 14:26:34 pve1 pmxcfs[771]: [dcdb] crit: cpg_initialize failed: 2
May 25 14:26:34 pve1 pmxcfs[771]: [dcdb] crit: can't initialize service
May 25 14:26:34 pve1 pmxcfs[771]: [status] crit: cpg_initialize failed: 2
May 25 14:26:34 pve1 pmxcfs[771]: [status] crit: can't initialize service
May 25 14:27:56 pve1 pvescheduler[1547]: jobs: cfs-lock 'file-jobs_cfg' error: got lock request timeout
May 25 14:28:09 pve1 pvescheduler[1593]: jobs: cfs-lock 'file-jobs_cfg' error: got lock request timeout

Additional Information:

• My status page at status.zayy.pro shows other services as down, but the hosted stuff should be online (excluding the Proxmox domain, which uses Cloudflare tunnels for login).

Request:

1. Troubleshoot Uptime Issues:
   • Any guidance on diagnosing and resolving the frequent server disconnects would be greatly appreciated.
2. Scheduled Server Reboots:
   • I’d like to set up an automatic weekly reboot for both servers. Ideally, this should occur every Monday at 5 AM EST.

Hardware information:

• I have 2 mini pcs, that are very recommended, and I am running very easy task, so it should not be crashing this much at least.

Thank you for your assistance!

 

[Maximiliano]

Hello,

The HA feature fences (reboots) nodes when they have issues, so the guests can be migrated over without risking data corruption. In a two node cluster it means that every time connection is lost for over a minute the entire cluster will reboot. We recommend using a QDevice to avoid such issues. You can read more about QDevices at [1] and about how fencing works at [2].

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support
[2] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#ha_manager_fencing

 

[Zayy]

Hello,

The HA feature fences (reboots) nodes when they have issues, so the guests can be migrated over without risking data corruption. In a two node cluster it means that every time connection is lost for over a minute the entire cluster will reboot. We recommend using a QDevice to avoid such issues. You can read more about QDevices at [1] and about how fencing works at [2].

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_corosync_external_vote_support
[2] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#ha_manager_fencing

Hey! I'm confused on how to setup [1], is there some video?