Hi everyone, I’ve been working on my homelab for a year and a half now, and I’ve tested several approaches to managing NAS and selfhosted applications. My current setup is an old desktop computer that boots into Proxmox, which has two VMs:
- TrueNAS Scale: manages storage, shares and replication.
- Debian 12 w/ docker: for all of my selfhosted applications.
The applications connect to the TrueNAS’ storage via NFS. I have two identical HDDs as a mirror, another one that has no failsafe (but it’s fine, because the data it contains is non-critical), and an external HDD that I want to use for replication, or some other use I still haven’t decided.
Now, the issue is the following. I’ve noticed that TrueNAS complains that the HDDs are Unhealthy and has complained about checksum errors. It also turns out that it can’t run S.M.A.R.T. checks, because instead of using an HBA, I’m directly passing the entire HDDs by ID to the VM. I’ve read recently that it’s discouraged to pass virtualized disks to TrueNAS, as data corruption can occur. And lately I was having trouble with a selfhosted instance of gitea, where data (apparently) got corrupted, and git was throwing errors when you tried to fetch or pull. I don’t know if this is related or not.
Now the thing is, I have a very limited budget, so I’m not keen on buying a dedicated HBA just out of a hunch. Is it really needed?
I mean, I know I could run TrueNAS directly, instead of using Proxmox, but I’ve found TrueNAS to be a pretty crappy Hypervisor (IMHO) in the past.
My main goal is to be able to manage the data that is used in selfhosted applications separately. For example, I want to be able to access Nextcloud’s files, even if the docker instance is broken. But maybe this is just an irrational fear, and I should instead backup the entire docker instances and hope for the best, or maybe I’m just misunderstanding how this works.
In any case, I have some data that I want to store and want to reliably archive, and I don’t want the docker apps to have too much control over it. That’s why I went with the current approach. It has also allowed for very granular control. But it’s also a bit more cumbersome, as everytime I want to selfhost a new app, I need to configure datasets, permissions and mounting of NFS shares.
Is there a simpler approach to all this? Or should I just buy an HBA and continue with things as they are? If so, which one should I buy (considering a very limited budget)?
I’m thankful for any advice you can give and for your time. Have a nice day!
Proxmox supports ZFS natively with management in the WebUI. So you could get rid of TrueNAS entirely and not need to deal with HBA pass-through or anything.
You also wouldn’t need NFS or have to deal with shares, as the data is available directly to Proxmox Containers via bind mounts.
Okay, if Proxmox can handle all that, I’ll be glad to ditch TrueNAS. However, I’m afraid that I won’t know how to migrate. I’ve found this reddit thread about someone who tried to do the same thing (I think) and accidentally corrupted their pools. About skipping NFS shares, that would be a big improvement for me, but I’m very unfamiliar with bind mounts. If I understand correctly, you can specify directories that live on the Proxmox Host, and they appear inside the VM, right? How does this compare to using virtual storage? Also, how can I replicate the ZFS pools to an external machine? In any case, thank you for that info!
Migration should be as simple as importing the existing ZFS stuff into the Proxmox OS. Having backups of important data is critical though.
If I understand correctly, you can specify directories that live on the Proxmox Host, and they appear inside the VM, right?
Inside a Container, yep. VMs can’t do bind mounts, and would need to use NFS again to share existing data from the host to inside the VM.
How does this compare to using virtual storage?
Like a VM virtual disk? Those are exclusive to each VM and can’t be shared, so if you want multiple VMs to access the same data then NFS would be needed.
But containers with bind mounts don’t have that limitation and multiple containers can access the same data (such as media).
Also, how can I replicate the ZFS pools to an external machine?
ZFS replication would do that.
Debian supports zfs so why the extra hassle of truenas? Seems to be a lot of extra work to add a vm when you could just use zfs in Debian. Or install something like Minio in Debian and use that to manage the data in s3 style buckets, again no vm needed
Okay, no VM, understood, but I do want to use a GUI for ZFS, because I’m basically a noob.
One thing to consider is that once zfs is setup there really is no significant intervention that is needed. I probably haven’t done anything to my proxmox zfs array in years.
I know its almost a meme to say just learn command line, but unfortunately in this case it will really help you understand what is happening and it also just takes a few commands to setup up once and then never worry about it again.
After the inital setup, the zfs GUI will be pretty much unused.
For commands I don’t use often, i use a note taking software to keep track of commands I used during setup because years go by before I use it again. I find the GUI often changes in that time making it harder to replicate whereas command line is the same and easier to document.
Well erm that I can’t get around… but I guess it depends on what you want to self host. I use truenas as a storage solution, standalone and it works very well for that - I have had it up and running for years and years. It also stores personal files and is not exposed to the internet apart from updates and a few other things. Then I have an old 4-node server from Supermicro which hosts the stuff I need and uses truenas to store the data. Now that is probably overkill but the data is safe and backed up well. Truenas has apps for a lot of things, and other ways to host docker containers if you have the latest scale community edition. So perhaps you could do it that way
I have two identical HDDs as a mirror, another one that has no failsafe (but it’s fine, because the data it contains is non-critical)
On separate pools, I hope? My understanding of ZFS is that the loss of any vdev will mean the loss of the pool, so your striped vdev should be in its own pool that you don’t mind losing.
Correct, they’re two separate pools, and the stripe one doesn’t contain any valuable data.
I hope you have backups
Never should you pass disks though by VM. The disk and SMART data is not passed though so ZFS has no way of knowing disk status. It also has more overhead than just passing though a PCIe device.
I would strongly recommend that you fix this ASAP
Yes, that’s why I’ve posted this question, and I immediately powered the entire NAS off, as to avoid any damage. It’s currently still powered off, until I find the best way to move forward. What I’m afraid of is that if I try to import the pools that were managed by the TrueNAS VM into a bare metal TrueNAS install, or Proxmox, that it won’t work correctly, or that I could lose data.
I would just pickup a simple PCIe sata card. They are fairly inexpensive and are idea for what you are doing.
Going forward, make sure you keep backups of important data.
I would probably shy away from passing the raw disk. There are a few dozen ways to skin that, But in the end I would probably just mount the disc through NFS,smb, whatever it takes. Reading that smart data is paramount for your situation. You could have a bad cable and never know it.
You could run a couple of VMs K8S and longhorn the data, It’s capable of backing up to an S3 compliant storage.
For my home stuff at the moment I’m running unraid with a BTRFS and a parity disc. The first every month I run a scrub if I had any corruption it would find it for me and alert me. It’s slow as balls but more protection is better than less. You can also buy some recycled discs and keep a offline store. I don’t love the recycled discs but for backup purposes that aren’t running 24x7 they’re better than nothing.
The problem is that I’ve never worked with Kubernetes, and it wouldn’t allow me to port over my current docker instances, afaik. And I’m not ready to pay for Unraid, I’m sorry, I need to set up the server as frugally as possible. Still, thank you for the info, I’ll keep it in mind, while I decide how to move forward.
I’ll keep it in mind, while I decide how to move forward. How I do it isn’t necessarily ideal. just throwing out options.
The chances of corruption from passing virtualized disks in truenas is lower than passing raw disks. I’d move them to virtualized disks.
As far as your SMART data, make a ventoy+debian live usb, boot your host off that and install smartmontools and use smartctl to read your SMART data. You host should still have access to that data. If you have smart errors, you’ll need to fix those first.
You should have all your data separately stored, it shouldn’t be locked inside containers, and using a VM hosted on a device to serve the data is a little convoluted
I personally don’t like TrueNAS - I’m not a hater, it just doesn’t float my boat (but I suspect someone will rage-downvote me 😉)
So, as an alternative approach, have a look at OpenMediaVault
It’s basically a Debian based NAS designed for DIY systems, which serves the local drives but it also has docker on, so feels like it might be a better fit for you.
I tried OMV in the past, but I found TrueNAS to be more intuitive… but that’s just personal preference I guess, and I’m not opposed to using OMV. Are you suggesting, then, that I run OMV on bare metal, and use it for everything? Or should it be inside a VM? If it’s the former, how easy is it to setup docker, because I’m not that familiar with OMV (it’s been a long time since I last checked it out). Is it like installing it in Debian directly? How does it handle the storage?
I always prefer bare metal for the core NAS functionality. There’s no benefit in adding a hypervisor layer just to create an NFS / SMB / iSCSI share
OMV comes with it’s own bare metal installer, based on Debian, so it’s as stable as a rock.
If you’ve used it before, you’re probably aware that it needs it’s own drive to install on, then everything else is the bulk storage pool… I’ve used various USB / mSATA / M.2 drives over the years and found it’s a really good way to segregate things.
I stopped using OMV when - IMO - “core” functions I was using (ie syncthing) became containers, because I have no use for that level of abstraction (but it’s less work for the OMV dev to maintain addons, so fair enough)
So, you don’t have to install docker, OMV automatically handles it for you.
How much OMV’s moved on, I don’t know, but I thought it would simplify your setup.
I run a docker host in Proxmox using ZFS datasets for the VM storage for things like my mailserver and NexcloudAIO. When I backup the docker VM, it snapshots the VM at a point in time, and backs up the snapshot to PBS. I’ve restored from that backup and it’s like the machine had just shut down as far as the data is concerned. It journals itself back to a consistent state and no data loss.
I wouldn’t run TrueNAS at all because I have no idea how that’s managing it’s storage and wouldn’t trust the result.
Wait, so if I understood correctly, you’re managing the ZFS pools directly in Proxmox, and then you have a VM that’s running docker, and using the storage that is managed by Proxmox, right? Hmm, sounds like a good solution. Is there any documentation or article that you could recommend, so that I can take a closer look? Also, how could I handle SMB shares?
Yes. So my debian docker host has some datasets attached:
mounted via fstab:
and I specify that path as the datadir for NCAIO:
Then when PBS calls a backup of that VM, all the datasets that Proxmox is managing for that backup take a snapshot, and that’s what’s backed up to PBS. Since it’s a snapshot, I can backup hourly if I want, and PBS dedups so the backups aren’t using a lot of space.
Other docker containers might have a mount that’s used as a bind mount inside the compose.yml to supply data storage.
Also, I have more than one backup job running on PBS so I have multiple backups, including on removable USB drives that I swap out (I restart the PBS server to change drives so it automounts the ZFS volumes on those removable drives and is ready for the next backup).
You could mount ZFS datasets you create in Proxmox as SMB shares in a sharing VM, and it would be handled the same.
As for documentation, I’ve never really seen any done this way but it seems to work. I’ve done restores of entire container stacks this way, as well as walked the backups to individually restore files from PBS.
If you try it and have any questions, ping me.
@thelemonalex I usually find that sata cable or connector are bad.
Personally I use proxmox as a host and share bulky nfs mounts for each VM like Immich, Plex/Jellyfin. For gitea and other small VM - use VM virtual drive and back it up periodically
So if I understand correctly, you’re managing the storage directly within Proxmox, instead of using a VM for that, right? Are the tools good? Does it support ZFS replication, SMB and things like that? Edit: I’ll also check the SATA cables, thanks!
@thelemonalex I use native linux nfs server. Manage shares using /etc/exports file.
For SMB, I use vm with samba and mount a few folders over nfs from proxmox (it just gives me a static IP in case if I decide to change something from proxmox side).
I manage zfs on the proxmox host. I think you can’t export zfs dataset to be managed by VM (except drive passthrough).
My infra has 1 proxmox bulky node with 2 HDD (zfs) and 3 nodes with small ssd/nvme in the cluster.
I was in somewhat of a similar situation but some of the SATA ports were bad and ended up buying a cheap SATA to PCIE card and passing it instead but that still had some issues so I ended up moving TrueNAS to its own dedicated machine and haven’t had any problems since.
Just my two cents.
To anyone reading, do NOT get a PCIe SATA card. Everything on the market is absolute crap that will make your life miserable.
Instead, get a used PCIe SAS card, preferably based on LSi. These should run about $50, and you may (depending on the model) need a $20 cable to connect it to SATA devices.
I have a cheap PCIe card I bought and it works fine.
It cost like $15 and has been rock solid. What is the issue?
The one I had would frequently drop the drives, wreaking havoc on my (software) RAID5. I later found out that it was splitting 2 ports into 4 in a way that completely broke spec.
Are there any specific limitations/requirements? Any recommended models or things to look out for? I looked on Amazon, and they range from around $30 to $200, and I really have no criteria, and I want to spend as little as possible.
If it’s an LSI card then make sure it’s either been flashed into IT mode, is capable of being flashed into IT mode, or is relatively modern and has that option built in.
What you really want is an HBA, but HBAs can be expensive, a raid card flashed to act as an HBA is typically much cheaper. A 6 gbit SAS card will do 3gbit sata, and no hard drive should be writing more than 3gbit. If you want to do SSDs then find a relatively more modern 12 gbit SAS card which will do 6gb sata.
I guess also look out for the REALLY old ones that won’t do over like 3tb. But I bought one of those for $20 almost 10 years ago so that shouldn’t be a concern. Those are probably all in the trash by now.
Thank you very much for the info. In the case of Raid Cards that can be flashed, is there something I need to look out for, besides the speed? Only HDDs will be used, so speed isn’t a priority.
I don’t want to speak to your specific use case, as it’s outside of my wheelhouse. My main point was that SATA cards are a problem.
As for LSi SAS cards, there’s a lot of details that probably don’t (but could) matter to you. PCIe generation, connectors, lanes, etc. There are threads on various other homelab forums, truenas, unraid, etc. Some models (like the 9212-4i4e, meaning it has 4 internal and 4 external lanes) have native SATA ports that are convenient, but most will have a SAS connector or two. You’d need a matching (forward) breakout cable to connect to SATA. Note that there are several common connectors, with internal and external versions of each.
You can use the external connectors (e.g. SFF-8088) as long as you have a matching (e.g. SFF-8088 SAS-SATA) breakout cable, and are willing to route the cable accordingly. Internal connectors are simpler, but might be in lower supply.
If you just need a simple controller card to handle a few drives without major speed concerns, and it will not be the boot drive, here are the things you need to watch for:
- MUST be LSi, but it can be rebranded LSi. This includes certain cards from Dell and IBM, but not all.
- Must support Initiator Target (IT) mode. The alternative is Initiator RAID (IR) mode. This is nearly all, since most can be flashed to IT mode regardless
- Watch for counterfeits! There are a bunch of these out there. My best advice is to find IT recyclers on eBay. These cards are a dime a dozen in old, decommissioned servers. They’re eager to sell them to whomever wants them.
Also, make sure you can point a fan at it. They’re designed for rackmount server chassis, so desktop-style cases don’t usually have the airflow needed.
Thank you very much for the detailed information. I’ll look into eBay again, maybe I can find a good offer that works. I’m unsure how to choose the number of lanes. Does that relate to the number of drives it supports? Also, in terms of cooling, would any PC case fan be enough, if strapped onto the HBA?
Kind of. They will be multiples of 4. Let’s say you got a gigantic 8i8e card, albeit unlikely. That would (probably) have 2 internal and 2 external SAS connectors. Your standard breakout cables will split each one into 4 SATA cables (up to 16 SATA ports if you used all 4 SAS ports and breakout cables), each running at full (SAS) speed.
But what if you were running an enterprise file server with a hundred drives, as many of these once were? You can’t cram dozens of these cards into a server, there aren’t enough PCIe slots/lanes. Well, there are SAS expansion cards, which basically act as a splitter. They will share those 4 lanes, potentially creating a bottleneck. But this is where SAS and SATA speeds differ- these are SAS lanes, which are (probably) double what SATA can do. So with expanders, you could attach 8 SATA drives to every 4 SAS lanes and still run at full speed. And if you need capacity more than speed, expanders allow you to split those 4 lanes to 24 drives. These are typically built into the drive backplane/DAS.
As for the fan, just about anything will do. The chip/heatsink gets hot, but is limited to the ~75 watts provided by the PCIe bus. I just have an old 80 or 90mm fan pointing at it.
Thank you, that’s actually very helpful, because I’m also kinda concerned that the HBA (or SATA to PCIE card) solution won’t fix this issue entirely, and it’s also relatively old system, and I’m worried about compatibility. Having two machines though is also kind of out of the question, because I don’t want to use that much electricity, it’s pretty expensive here where I live. I do have a Raspberry Pi, which could be used for some selfhosted services, but not the demanding ones, and not running alongside each other. I mean, my current PC sometimes struggles already, so I guess the Raspberry Pi will perform even worse… Maybe I can find a solution that allows to still manage storage reliably and flexibly, and also be able to have docker containers.
TrueNAS Scale is the Linux version of TrueNAS with docker support. Might be worth looking into. It also supports doing an in place upgrade from Core but you wont be able to revert the same way.
I’m already running TrueNAS Scale, but I’m unsure about the docker support. I did use TrueCharts and such, but it was very unstable, so I moved away from that.
If you’re only doing a VM or two, I’d get rid of proxmox and run truenas directly. It’s gotten better for VMs.
Also make sure you read up on the ecc requirements for truenas if you’re not using ecc ram
And is it easy for the docker instances inside the VM to access the host’s datasets? About ECC, thank you for bringing it up, because I actually have no idea on the subject, and I’m sure that my current ram isn’t ECC. I’ll look into it. It could explain the issue I had with gitea, right?
If you pass a whole raw disk, not virtualized, then TrueNAS should not complain. I don’t know if you can do that in proxmox, I haven’t tried.
Personally I’d get rid of TrueNAS. Even if docker is down, the VM with the data is still up and accessible over anything running on the VM, like scp via ssh.
If you pass a whole raw disk, not virtualized, then TrueNAS should not complain.
Although Smart data counters live on each individual disk, it is accessed via the disk controller interface. No controller, no smart data.
If you pass the disk the meta data isn’t passed though including SMART data so it is a ticking time bomb.
I passed the whole raw disk in theory, but I’ve read that even if you do that, it’s still virtualized somehow. But I don’t know enough about it, I’m just basing this on the research I’ve done so far. About skipping TrueNAS, maybe you’re right, but if I also want to set up SMB shares? Is there a tool or something that makes it easy for someone with limited experience? It would also need to handle replication of data, and the advantage of ZFS Replication over using something like rsync, is that you can simply push the delta between snapshots, instead of replacing entire files, or at least that’s how I’ve understood this.
Send your question to the podcasters at 2.5admins, this seems right up their alley.
Thank you, I might raise the issue there, if I struggle to fix the issue. I didn’t know the podcasters, but now I’ll try and listen to a few episodes, and maybe I can continue learning. Thank you for the suggestion
They’re gonna say the same thing you’ve read here, which is that if you’re going to virtualize TrueNAS, pass through the controller, not just the disks.
They always do Free consulting on air, so if your issue gets to them and they have a spot, you get a very very competant answer
