• Simple title, simple question. As my big post in the General area of the board was maybe a bit too broad and too much, I want to go a bit more into details in different threads.


    I am planning to get started with OMV and use currently BTRFS on a Synology NAS. One of the biggest inital decision after Hardware and OS is of course the FS. Since I want Bitrot detection and good data integrity, I think the main two options for me would be BTRFS or ZFS.


    However, I get the feeling that both are not a simple choice for OMV. First one is officially supported, but it's mentioned that a lot of features don't work out of the box and need manual tweeking. ZFS however is a BSD FS and not officially supported. On top there seems to be strong considerations to switch completly to BTRFS in the future https://github.com/openmediavault/openmediavault/issues/101 .


    So of course that leaves me a bit puzzled. And of course there are a lot of consecutive questions already in my mind, which I think are only worth diving into as soon as I have a strong candidate for a FS on OMV.

    • Offizieller Beitrag

    ZFS however is a BSD FS and not officially supported

    zfs started as a Solaris filesystem. It was ported to FreeBSD. But, FreeBSD is actually moving to the ZFS on Linux codebase. Not sure what officially supported means but it is supported by Canonical and Debian. Just because it isn't in the kernel doesn't mean it isn't supported. It is considered production ready.


    btrfs filesystems can be mounted in the OMV web interface but most btrfs features have to be configured from the command line. And it won't be the only filesystem in later versions of OMV because plugins can be written.


    As for which one to use, that is personal preference. Try them both. Don't rule out snapraid if you just want bitrot protection.

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • With supported I referred to this: https://openmediavault.readthe…/storage/filesystems.html
    Hmm it's hard to try them out right now, since I don't have the needed hardware already. Also I am not really sure how to test them out. Do you mean configuration wise? It might be a bit complicated to provoke data threatening scenarios.


    Does anybody have a link to a got comparison or decision making guide for this question?


    Never heard of snapraid. I firstly had BTRFS in mind, because this was the initial solution I stumbled over that could help with data integrity and that was available on Synology.


    I am also not super sure whether my hardware plans may be important for the decision. I tend to start with two new WD red 12 or 10 TB disks and connect them depending on the board directly or via oculink or similar. In case they should get full I would like to get some more and ideally integrate them to the existing volume. Synology have their own raid system to do this. I don't know whether this is possible with one of the three FS under OMV.


    I am not a big FS expert, which makes things maybe not easyer. So far I just tended to take the recommended one at the time and task, like ext3, ext4, fat, ntfs and whatsoever.

    • Offizieller Beitrag

    With supported I referred to this

    That is referring to core support. Plugins can add support.


    Hmm it's hard to try them out right now, since I don't have the needed hardware already. Also I am not really sure how to test them out. Do you mean configuration wise? It might be a bit complicated to provoke data threatening scenarios.

    Try it in a VM. It is easy to remove a drive from a VM to simulate a drive failure.


    Does anybody have a link to a got comparison or decision making guide for this question?

    That would be difficult to create since the decision needs to be based on your skills and comfort level. I would not put data on a filesystem you don't understand or know how to fix. This comparison is good - https://www.snapraid.it/compare


    In case they should get full I would like to get some more and ideally integrate them to the existing volume.

    This is a very common situation with OMV and many people are using mergerfs (unionfilesystem plugin) and snapraid. mergerfs allows you to pool the drives while making it super easy to expand the pool. Each drive can run ext4 that many are comfortable with. Snapraid gives you the bitrot protection and redundancy. No matter which solution you pick, don't forget backup.

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

  • Thanks for the link Morlan.


    I am not really experienced with VM either. I only tried once to set up a Windows 10VM with gnome boxes because of some stupid xiaomi software, but that never booted properly after installation.


    In principle you are right regarding knowledge about FS. In fact I believe that most problems people have with IT is that they are not really knowing what they are doing. However this is a tough thing as depending on how deeply one takes this, it's a tough task to fully understand a FS. I would say I have a rough general understanding and I understand the idea of using has sums for data integrity. Maybe I will try to read a bit more about ext4. And the possibilities for OMV. In fact with my Diskstation I once had to repair a carshed ext4 raid1. That was of course frightening.


    With Backups I have no worries I would say. Currently I do daily encrypted backups to gdrive and have a manually triggered backup to usb. Currently with the Synology backup software but maybe I will install duplicati via docker on the DS and try to create a backup to GDrive in advance. I have some more questions regarding to this, but this is maybe the wrong place,

  • I read the Article that Morlan linked. And eventually I really consider to use this system. There are two big differences from my use and features I am aware of from BTRFS. First it is not really a raid which has ups and downs. A plus for BTRFS is that I have instantaneous parity, which creates a window where things could happen to the file and I can't recover it. Second thing is that with BTRFS I can create tremandous amounts of snapshots that don't really take up much of extra space provding some kind of version control. I activated this on my NAS, but never really made use out of it. As far as I understood snapRaid is lacking this.
    Some things that are unclear to me are: If I create a file in the middle of a window. Will it be stored together with a hash directly so that when the sync occurs I could get a feedback whether the file went corrupt during the window? I am afraid it's not the case. If so, one would know directly after the next sync attempt that the file is messed up. Second the author says that there is indeed some kind of file level restore function once the files are synced. But does it only cover the possibility to go back to the former version before the next sync happened? My understanding would be the following: I have a file called myfile.odt that was synced already and is present on sda1 and sdb1. Now I do some things to the file and save it. MergerFS will store it eg to sda1, leaving the old version on sdb1. Now I could get back to the old version, because it is still present on sdb1. But after the next sync both files are identical again and I can't recover the old version from my volume. Is that correct?


    In addition it was interesting, because I became aware of ansible, which might be handy. I maintain three different machines with arch-linux and for initial setup I have a nearly identical script for installation. Might be possible to automate this in an easy way.


    This seems also to be handy: https://michaelxander.com/diy-nas/ At least it solved the puzzle how snapRAID and MergerFS work together and that not the role of data and parity is mixed.


    Now I have to consider if I can live with this danger window hmm.

  • Maybe another question regarding BTRFS. Is there an overview what features are lacking in OMV and what would be need to configured manually?
    I am not afraid of the command line. And I assume that of course the command line work is refering to initial install and configuration and theres not the need for constant sshing to the machine. I would be espacially interested how the adding of disks to the BTRFS volume would work under OMV.
    I also read the recommendation that one would need to use a different Kernel.

    • Offizieller Beitrag

    Is there an overview what features are lacking in OMV and what would be need to configured manually?

    It is easier to mention what features OMV has for btrfs. Basically you can create single drive, simple btrfs filesystems and mount a btrfs filesystem. Everything else (raid, compression, sub-volumes, etc) needs to be configured from command line.

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

    • Offizieller Beitrag

    Maybe another question regarding BTRFS. Is there an overview what features are lacking in OMV and what would be need to configured manually?

    btrfs maintenance tools:
    https://github.com/kdave/btrfsmaintenance


    I appreciate snapper to create automatic snapshots of the OS.

  • Oh man, that looks really like a lot of manual work. Would it be different if one installs OMV on top of debian instead of a direct OMV installation? I read about this possibility but I still don't know what the benefit would be. Maybe stuff like this is better supported?

  • I don't know whtether I really need btrbk. What I would expect as a minimum for it to be usefull is creation of BTRFS Raids and the possibility to extend them to somethin like Raid 10. The Bitrot detection should work easily and scrubbing and repair should be automated. Snapshots would be nice, but it's not my main concern. With incremental backups it's not to mandatory and more like a nice to have. And last but really important, stability. Since I read something about custom Kernels etc. I am a bit afraid that maybe after some Kernel update the FS is getting messed up. I think something like this should be rock solid.


    Wnapraid seems to be the easiest solution together with merge fs. As far as i understood one could in principle with a start of two disks live without mergefs and without any problems if one wants to extend the volume configure it later on, as snapraid and mergefs are indepenedent from each other if used together. Although it might be less effort later on and it seems that it doesn't matter if one creates a mergefs volume with only one disk in the beginning.


    I am still highly undecisioned.

    • Offizieller Beitrag

    zfs will cache using ram and it makes zfs look like it is ram hungry but it will work fine without the ram (unless you have dedupe enabled).

    omv 7.0.4-2 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.10 | compose 7.1.2 | k8s 7.0-6 | cputemp 7.0 | mergerfs 7.0.3


    omv-extras.org plugins source code and issue tracker - github


    Please try ctrl-shift-R and read this before posting a question.

    Please put your OMV system details in your signature.
    Please don't PM for support... Too many PMs!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!