Shared storage live migration

JMMOCO

New Member
Sep 14, 2024
2
2
3
Good day,

I am looking for some insight into a lab setup of Proxmox VE 8.3 I am testing with. The goal would be HA with failover. The setup is as follows:

2 servers with SAS controllers (1 Linux workstation as a QDevice to meet the requirement of 3)
1 SAN with redundant SAS connections
Multipath setup for 2 connections from redundant controllers on SAN
LVM storage configuration

I have several servers running without issue but when I try to migrate the VM from one host to the other, it seems to kick off a copy of data which seems illogical since the entire VM should be running within the shared storage container. Can anyone shed some light on why migrating compute seems to also force some sort of storage data migration as well? The disks seem to be recreated with new numbers at end because it sees the originals and conflicts, creating more drives with (3), (4), (5) etc. at the end.

As I am new to this setup, I am looking for thoughts, ideas, best practices and suggestions. Any info is appreciated. Thanks.
 
Hi @JMMOCO , welcome to the forum.

Can you provide :
- output of "cat /etc/pve/storage.cfg"
- output of "pvesm status"
- output of : pvs;vgs;lvs
- output of "qm config [id]"

The issue is likely due to the storage pool not being marked as "shared".


Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox
Hi BBGEEK17,

Well, that was exactly what it was. Apparently, after configuring multipath, I created the physical volume and volume group through CLI and must have missed a switch or something because "Shared" was not ticked. I'm not sure if this had to be configured after the fact in the GUI but once I checked the box, all works as expected. I appreciate the welcome and the quick and knowledgeable advice. I look forward to hanging around and learning.

Thanks again.