[GH-ISSUE #37] High Availability (HA) #28

New issue

Closed

opened 2026-03-02 15:47:20 +03:00 by kerem · 3 comments

kerem commented

2026-03-02 15:47:20 +03:00

Owner

Originally created by @bernuchonsebastien on GitHub (Feb 3, 2026).
Original GitHub issue: https://github.com/PegaProx/project-pegaprox/issues/37

Hello,

In the settings, I have High Availability (HA) enabled, and I also have Proxmox Native HA enabled.
Does this cause a conflict? What are the best practices?

Thanks your respond
Seb

Originally created by @bernuchonsebastien on GitHub (Feb 3, 2026). Original GitHub issue: https://github.com/PegaProx/project-pegaprox/issues/37 Hello, In the settings, I have High Availability (HA) enabled, and I also have Proxmox Native HA enabled. Does this cause a conflict? What are the best practices? Thanks your respond Seb

kerem closed this issue

2026-03-02 15:47:21 +03:00

kerem commented

2026-03-02 15:47:22 +03:00

Author

Owner

@mkellermann97 commented on GitHub (Feb 3, 2026):

Hello @bernuchonsebastien ,

this honestly depends on how you are using Proxmox native HA. If you are already implemented and are using Proxmox native HA we wouldn't recommend using also PegaProx ones. As this can lead to conflicts.

Regards,
Marcus

@mkellermann97 commented on GitHub (Feb 3, 2026): Hello @bernuchonsebastien , this honestly depends on how you are using Proxmox native HA. If you are already implemented and are using Proxmox native HA we wouldn't recommend using also PegaProx ones. As this can lead to conflicts. Regards, Marcus

kerem commented

2026-03-02 15:47:22 +03:00

Author

Owner

@bernuchonsebastien commented on GitHub (Feb 4, 2026):

Thanks Marcus,

We have a cluster with 4 nodes and quorum running on Ceph.
We were using the native High Availability (HA) for VM failover.
Is it possible to disable HA at the VM level and use Pegaprox’s HA instead, without impacting the Ceph cluster?

Regards,
Seb

@bernuchonsebastien commented on GitHub (Feb 4, 2026): Thanks Marcus, We have a cluster with 4 nodes and quorum running on Ceph. We were using the native High Availability (HA) for VM failover. Is it possible to disable HA at the VM level and use Pegaprox’s HA instead, without impacting the Ceph cluster? Regards, Seb

kerem commented

2026-03-02 15:47:22 +03:00

Author

Owner

@mkellermann97 commented on GitHub (Feb 4, 2026):

Hi @bernuchonsebastien ,
Yes, this is absolutely possible and won't impact your Ceph cluster at all.
Ceph and Proxmox HA are independent layers – Ceph maintains its own quorum through the MON daemons, so disabling Proxmox HA at the VM level has no effect on your storage cluster as far as we know.
Here's how to switch:
Disable Proxmox Native HA for each VM
via PegaProx GUI: Datacenter → Native Proxmox HA → Resources → Remove
Enable PegaProx HA:
In PegaProx, go to Settings → HA → Enable High Availability -> Install Fencing Agent

That's it. PegaProx will then monitor your nodes every 10 seconds and automatically restart VMs on surviving nodes if a node fails around 1:30 min downtime.
A few things to keep in mind:
Since you're running Ceph, you're already on shared storage, so all VMs are eligible for HA failover

With 4 nodes, quorum works normally – no special configuration needed

PegaProx HA and Proxmox HA should not run simultaneously for the same VMs to avoid conflicts

Regards,
Marcus

@mkellermann97 commented on GitHub (Feb 4, 2026): Hi @bernuchonsebastien , Yes, this is absolutely possible and won't impact your Ceph cluster at all. Ceph and Proxmox HA are independent layers – Ceph maintains its own quorum through the MON daemons, so disabling Proxmox HA at the VM level has no effect on your storage cluster as far as we know. Here's how to switch: Disable Proxmox Native HA for each VM via PegaProx GUI: Datacenter → Native Proxmox HA → Resources → Remove Enable PegaProx HA: In PegaProx, go to Settings → HA → Enable High Availability -> Install Fencing Agent That's it. PegaProx will then monitor your nodes every 10 seconds and automatically restart VMs on surviving nodes if a node fails around 1:30 min downtime. A few things to keep in mind: Since you're running Ceph, you're already on shared storage, so all VMs are eligible for HA failover With 4 nodes, quorum works normally – no special configuration needed PegaProx HA and Proxmox HA should not run simultaneously for the same VMs to avoid conflicts Regards, Marcus