Hello!
I am relatively new to proxmox and ended up using PVE after VMware made it economically unfeasible to virtualize smaller environments, where a single server was hosting just a handful of virtual machines.
One thing I've been missing since we started using PVE is the ability to simply reinstall VMware in case of issues (broken systems, etc.) while keeping the datastores untouched and just reconnecting the virtual machines afterward.
I wanted to achieve the same level of convenience with PVE, so I set up a test environment. I’d be very grateful to hear the opinions of more experienced users here (which, in my case, means all of you) on whether this is a proper way to handle such a scenario:
Setup:
Then, I restored only the /etc/pve/storage.cfg and the VM configuration files from /etc/pve/nodes/tprox/qemu-server/*.conf. After a reboot, the RAID5 volume appeared correctly in the datacenter tree, and the virtual machines were available and could be started.
Am I missing something here, or is it really that simple to perform a kind of "repair installation" if the OS and VM's storage are separated? Is this a propper way to go for small production environments with single hosts?
Thank you very much for your feedback!
I am relatively new to proxmox and ended up using PVE after VMware made it economically unfeasible to virtualize smaller environments, where a single server was hosting just a handful of virtual machines.
One thing I've been missing since we started using PVE is the ability to simply reinstall VMware in case of issues (broken systems, etc.) while keeping the datastores untouched and just reconnecting the virtual machines afterward.
I wanted to achieve the same level of convenience with PVE, so I set up a test environment. I’d be very grateful to hear the opinions of more experienced users here (which, in my case, means all of you) on whether this is a proper way to handle such a scenario:
Setup:
- HPE Proliant Server (any other should do as well)
- 1x RAID1 SATA HDDs for the OS
- 1x RAID5 SAS 15k HDDs for the virtual machines
- Raids are controller based, no soft-raid or ZFS
- Single Host, no clustering
- Installed PVE 8.4 on the RAID1 volume
- Added the RAID5 volume as an LVM-thin storage
- Created Windows-based virtual machines on the RAID5 volume
- Backed up the /etc/pve directory
Then, I restored only the /etc/pve/storage.cfg and the VM configuration files from /etc/pve/nodes/tprox/qemu-server/*.conf. After a reboot, the RAID5 volume appeared correctly in the datacenter tree, and the virtual machines were available and could be started.
Am I missing something here, or is it really that simple to perform a kind of "repair installation" if the OS and VM's storage are separated? Is this a propper way to go for small production environments with single hosts?
Thank you very much for your feedback!
