How to migrate while in HA?

[ballybob]

Is the proper way to migrate a vm to another node while in HA is to set the vm to be "ignored" state in HA, then migrate?
If I don't do this, it just keeps failing back to the primary.

 

[LnxBil]

If I don't do this, it just keeps failing back to the primary.

That is not the default, at least for me. Have you changed settings?
What is your Datacenter -> Options -> Cluster Resource Schedulung.

 

[Impact]

Please share

tail -n+0 /etc/pve/ha/{resources,groups}.cfg
qm config VMIDHERE --current

as well as the task/migration history/log.

 

[bbgeek17]

In addition to the above posts, you may want to carefully review the documentation on HA, specifically here:
https://pve.proxmox.com/pve-docs/chapter-ha-manager.html#ha_manager_groups

---

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

 

[ballybob]

That is not the default, at least for me. Have you changed settings?
What is your Datacenter -> Options -> Cluster Resource Schedulung.

Default.

I do have an HA Group, which is set to restricted: no, nofailback: no with the nodes in it and watchdog fencing.

 

[bbgeek17]

nofailback: no

Did you mean that nofailback is unchecked? Assuming so, that means that the HA will failback as described below:

nofailback: <boolean> (default = 0)
The CRM tries to run services on the node with the highest priority. If a node with higher priority comes online, the CRM migrates the service to that node. Enabling nofailback prevents that behavior.

As such, when you migrate the VM from higher priority node, it will be moved back by HA.

---

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

 

[ballybob]

Did you mean that nofailback is unchecked? Assuming so, that means that the HA will failback as described below:

nofailback: <boolean> (default = 0)
The CRM tries to run services on the node with the highest priority. If a node with higher priority comes online, the CRM migrates the service to that node. Enabling nofailback prevents that behavior.

As such, when you migrate the VM from higher priority node, it will be moved back by HA.

---

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

Is yes the preferred method?
Right now, I do expect it will fail back to PVE1 but i figured it wouldn't if I manually did it.