Greetings Proxmoxians!
I've got a Proxmox network question that I hope someone can answer as I can't seem to find a clue myself.
I've read through forums, best practices and the Admin Guide.
There are 4 servers in a cluster. However, Ceph is not going to be used in this deployment, because the servers are two pairs of matched hardware. So it's kind of like a cluster of two clusters?
The aim of clustering is just a single pane of glass, and to facilitate offline and/or live migration and possibly storage replicas (using Proxmox VE 6.0) between "like" server pairs for planned maintenance/etc. A bit of downtime for a VM is ok during planned migration or VM recovery, but a host's reliability or networking fault would be too long of a downtime to experience. So I'm hardening the local host configuration and networking side of things.
The requirement for a separate (and preferably redundant) cluster/corosync network is understood.
My confusion lies in which particular network is used during VM disk migration and/or storage replication, if the disk needs to be migrated over the network.
Perhaps it is my reading of it, but it's not incredibly clear to me from the networking section of the guide:
https://pve.proxmox.com/wiki/Separate_Cluster_Network
Is the VM disk migration network the separate "cluster/corosync" network defined in the /etc/pve/corosync.conf? Or is it the native bridge the cluster uses? ... or something else I've not considered? I've only got 1x 10GbE NIC per server (and 2x 1Gb each as well), so would like to make sure the 10GbE is used for the most critical bandwidth hogs in the event of a needed downtime.
Perhaps the question would be rephrased as: How do I separate VM migration and/or storage replica traffic from the critical cluster traffic?
Thanks in advance (and big congrats on VE 6.0 release!)
I've got a Proxmox network question that I hope someone can answer as I can't seem to find a clue myself.
I've read through forums, best practices and the Admin Guide.
There are 4 servers in a cluster. However, Ceph is not going to be used in this deployment, because the servers are two pairs of matched hardware. So it's kind of like a cluster of two clusters?
The aim of clustering is just a single pane of glass, and to facilitate offline and/or live migration and possibly storage replicas (using Proxmox VE 6.0) between "like" server pairs for planned maintenance/etc. A bit of downtime for a VM is ok during planned migration or VM recovery, but a host's reliability or networking fault would be too long of a downtime to experience. So I'm hardening the local host configuration and networking side of things.
The requirement for a separate (and preferably redundant) cluster/corosync network is understood.
My confusion lies in which particular network is used during VM disk migration and/or storage replication, if the disk needs to be migrated over the network.
Perhaps it is my reading of it, but it's not incredibly clear to me from the networking section of the guide:
https://pve.proxmox.com/wiki/Separate_Cluster_Network
Is the VM disk migration network the separate "cluster/corosync" network defined in the /etc/pve/corosync.conf? Or is it the native bridge the cluster uses? ... or something else I've not considered? I've only got 1x 10GbE NIC per server (and 2x 1Gb each as well), so would like to make sure the 10GbE is used for the most critical bandwidth hogs in the event of a needed downtime.
Perhaps the question would be rephrased as: How do I separate VM migration and/or storage replica traffic from the critical cluster traffic?
Thanks in advance (and big congrats on VE 6.0 release!)
