(Proxmox) Single drive for Boot and VMs
- Thread starter servape
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Why does placing root and VM disk on the same drive increase wear-out? Let's say compared to placing root and VM disk on each 256GB M.2?
I would claim it would be even worse.
I don't in general see a lot of IO of the root-volumes, regardless of being root-volume of VM or the hypervisor.
Last edited:
Wear out an SSD? From what I've read, most people use SSD for speed/performance of the VMs and OS.. So you put VMs on a standard mechanical HDD?
"screams consumer SSD" yes. This is for home use.
I like boot separate, with RAID1 even. When things go sideways for a node I want to have a bootable system for diagnosing. Homelab also. I like it when systems can still limp along even with 50% of components dead. To me resilience is a good virtue.
Putting VMs on same drive as root would hasten the decline of your boot drive. I would want to at least mirror in this situation.
Consumer SSDs have a fraction of the endurance of enterprise- or datacenter-oriented SSDs. They also lack things like power loss protection. Depending on your VM workloads, this may or may not be a problem.
You can get used datacenter SSDs for pennies on the dollar. See Buyer’s Guide to getting a used datacenter SSD inexpensively.
Whatever floats your boat. ;-) Recommend then to get a single Micron MAX 5100 SATA for cheap and install everything on it and be done again for 10 years. Drive has a large TBW rating, i.e. hard to kill with writes. Firmware update available from micron.com after registration. Power-loss protection. Low power usage.
This project was put on the side and I am back on it again. I am looking into used, enterprise SSDs like several have mentioned in this thread. I would like the Proxmox boot drive to be strickly the OS only and in RAIDZ1. Does anyone have any recommendations for a low cost - low capacity (128GB range) and reliable/high endurance drive?
I love discussions like this where everybody gets all pedantic and raves about things like SSD "wearout" without even considering the actual use case. You usually get this with the "ZFS requires ECC" thingy. But this one is a fun basket advice focusing in on the wrong things too.
Putting the boot and VMs together on one drive really is not a worrisome thing for homelab use cases. There is no actual data available to show otherwise.
For a home lab, running typical home lab things, and with modern (last two years) better consumer SSDs this is not a real worry. Look at the actual wear level performance specs and consider how much (or, really, how little) disk write things like Home Assistant or other home-lab apps do. Don't pull thing out of your used laptop or go for the cheapest Samsung EVO drive - stay to the middle/upper range of consumer SSDs and you are looking at YEARS of use.
Now - if the OP was asking about an NVR app for 10 cameras with continuous recording the story is quite different.
ps - there are good reasons not to do this too. But they really revolve around life cycle of your lab and the value of having boot and VMs separate when you are making changes or replacing things.
pps - if the pedants want to give some better advice they should really be nagging you to make sure you do regular backups. All drives actually do fail. Not because you mixed uses on them in ways that bother people. But because they are physical things that break. And all of the wear-level specs are really just statically derived. Most drives do much better than expected. Often 5-10x better. But some drives fail on day two. Do backups! Do backups!! Do backups!!!
Putting the boot and VMs together on one drive really is not a worrisome thing for homelab use cases. There is no actual data available to show otherwise.
For a home lab, running typical home lab things, and with modern (last two years) better consumer SSDs this is not a real worry. Look at the actual wear level performance specs and consider how much (or, really, how little) disk write things like Home Assistant or other home-lab apps do. Don't pull thing out of your used laptop or go for the cheapest Samsung EVO drive - stay to the middle/upper range of consumer SSDs and you are looking at YEARS of use.
Now - if the OP was asking about an NVR app for 10 cameras with continuous recording the story is quite different.
ps - there are good reasons not to do this too. But they really revolve around life cycle of your lab and the value of having boot and VMs separate when you are making changes or replacing things.
pps - if the pedants want to give some better advice they should really be nagging you to make sure you do regular backups. All drives actually do fail. Not because you mixed uses on them in ways that bother people. But because they are physical things that break. And all of the wear-level specs are really just statically derived. Most drives do much better than expected. Often 5-10x better. But some drives fail on day two. Do backups! Do backups!! Do backups!!!
The biggest downside I can think of is no redundancy, if the 500GB drive fails you lose all.
SSD wear will depend much on what you'll be runnning in your VM, but if not something write intensive like torrenting or DVR then there should be no problem with endurance.
SSD wear will depend much on what you'll be runnning in your VM, but if not something write intensive like torrenting or DVR then there should be no problem with endurance.
I think its always good practice to separate OS from Data. In this case Proxmox is the OS and the VM's the data.
If the root drive dies you have lost your VM's - and if you do not have a good backup strategy you are not in a good place.
If you have VM's on one drive, proxmox on the other, you have at least made it possible to reinstall proxmox and still have the data for your VM's available.
And also consider that proxmox when its running will not write a lot of data on your, but if you put VM's on the drive you are putting more stress onto the "important" drive for your VM solution.
It is certainly true that for a homelab many of the good practices does not matter much - but why not practice stuff that is considered best practice - even if its a hobby and you only have small requirements.
A disk can die at any time and if you put all your eggs in the same basket you are making it harder for your self.
Obviously you should also have a backup strategy in place.
So yes, anything goes, but you can buy a cheap small SATA ssd and put proxmox on that and would probably not have to worry much about it being worn out.
Small enterprise SSD's are cheap.
If the root drive dies you have lost your VM's - and if you do not have a good backup strategy you are not in a good place.
If you have VM's on one drive, proxmox on the other, you have at least made it possible to reinstall proxmox and still have the data for your VM's available.
And also consider that proxmox when its running will not write a lot of data on your, but if you put VM's on the drive you are putting more stress onto the "important" drive for your VM solution.
It is certainly true that for a homelab many of the good practices does not matter much - but why not practice stuff that is considered best practice - even if its a hobby and you only have small requirements.
A disk can die at any time and if you put all your eggs in the same basket you are making it harder for your self.
Obviously you should also have a backup strategy in place.
So yes, anything goes, but you can buy a cheap small SATA ssd and put proxmox on that and would probably not have to worry much about it being worn out.
Small enterprise SSD's are cheap.