Proxmox HA for instances is not working

[kiranch97]

Hi Team,

We have an query related to the Proxmox HA, we have 3 nodes setup with ceph as shared storage.
we have created a HA group under cluster and added VMs to the HA group.

Tested graceful rebooting one node in cluster, but the VM which was running on respective node was not migrate (either live migrate or restarted on another available node). VM is also part of the HA group.

What is happening is actually, when we shutdown node, all VMs on this physical node are autoshutting down and powered on after rebooted physical node comes online.

Can anyone help us on the Proxmox HA. Thank you for your help in advance.

 

[bbgeek17]

What do the following commands return in a healthy cluster?
ha-manager status
ha-manager config

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[kiranch97]

Hi @bbgeek17

 

[bbgeek17]

what about "ha-manager groupconfig"
You can also paste text output in the forum, using CODE tags.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[kiranch97]

 

[kiranch97]

Hi Team,

Can you please support on this issue? Thank you.

 

[bbgeek17]

Tested graceful rebooting one node in cluster,

how fast does the node come back from the reboot?
PVE clustering waits two+ minutes before moving the VM to a surviving node:
https://forum.proxmox.com/threads/p...dont-migrate-if-node-fails.73225/#post-327095

If your rebooted node came back sooner, then it may not have had a chance to move the VM. I suggest testing by completely powering down the node.
You can also run "journalctl -f" on the other two nodes prior to power down and watch the logs that may be helpful if the behavior you are seeing is by design, just not in line with your expectations.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[Khoi]

Hello.

I have the same problem. HA is not ok when I use graceful reboot but It worked if I press reset button on physical server. Boot time is same with these action.