Total NOOB with bubblewrap stupid questions

  • Hi guys, I am a total NOOB in every way to OMV.


    I have looked for some simple start guides, etc but none answer my questions. It seems you have to be a Linux expert.


    Once you have OMV set up...


    #1. If your hardware fails (Motherboard, computer, power supply, whatever) and you can't find a replacement hardware, and you NEED access to your OMV data, can you take one of your OMV drives out and plug it into a Windows machine with a USB SATA adapter, can you copy your data off the OMV drive to your windows computer? (Yes I know Windows is NTFS)


    #2. There isn't a straightforward guide that I can find that explains STEP BY STEP, if your OMV drive fails, the correct procedure to remove it, put a new one in, and OMV rebuilds the drive. (do you need to format the drive first? If so what format?)


    #3. If you want to expand storage space in the future, HOW is this done explained in a step by step method of removing the smaller drives and replacing with larger capacity drives.

    One at a time?


    My very first initial experience with OMV is very disapointment, because trying to ready through the forums, I am seeing conflicting info, sarcasm, squables, attitide.


    I am not seeing a welcoming warm environment, and I am hoping you can all show me otherwise.


    Just an FYI, I am not seeing much better on TrueNAS Scale for that matter.


    As my first OMV test, I just set up an HP T620 Thin client. I bought an M2 SSD to USB adapter and put a 1TB SSD in it, watched a few YouTube videos, and everything is working fine with it, however it just is a perfect single point of failure, and it was just as a test to see how OMV works and to get it on the network.


    I now want to use OMV with a true raid system, so I bought a few servers.


    HP Proliant ML310e Gen8 V2 tower,

    HP Proliant ML350 Gen8 tower (I love the look of these)

    Dell Poweredge T310 tower (two of them as they only use 75 Watts)

    Dell Poweredge 2900 tower (A real tank beast of a machine w/8 drive bays)


    Yes, 4 servers, overkill to the extreme.


    I am going to use some servers with OMV, and some with TrueNAS Scale.


    This will give me the ultimate redundancy, and I will also use two servers offsite to make everything bulletproof.


    I just wanted to explain the method to my madness.


    The CRAZY thing is... the simple single point of failure with the HP T-620 thin client and the USB M2 SSD...it's formatted as NTFS, which means I can grab that, plug it into a USB drive on any Windows machine, and have instant access to my data, which is my concern with question #1.


    Thank you all in advance. :)

  • crashtest

    Approved the thread.
  • Exact step by step instructions will be difficult to give because there are several different ways to achieve redundancy for your storage. There are different kinds of RAID and the process is different depending on your choice. There is also combination filesystem/raid setups like zfs and btrfs which are different again. You need to give more info and decide on what configuration you want before you will get step by step help.


    I personally use mdadm raid 5 with an xfs filesystem. Others use zfs raid z1 or z2, others use btrfs mirrors, while others run single drives and rsync them or put them in a mergerfs pool and run snapraid on top for redundancy.


    All that said I can answer your other question. If you want a properly functioning system you have to use a Linux filesystem (ext4, xfs, btrfs, zfs, etc.)


    Ntfs, fat32, Exfat, do not carry Linux ownership and permissions, which will cause things to misbehave and make you think it is all just a piece of junk.


    If you want to be able to unplug a drive and plug it into windows, you will have problems in 2 ways. Windows does not understand how to handle a Linux raid so if you use raid you can forget that idea. Windows also does not natively read/write a Linux filesystem the way Linux can read/write a windows filesystem. However there are filesystem drivers available from Paragon that will let windows use a Linux drive.

    Edited 2 times, last by BernH: fixing wrong spell check corrections ().

  • Hi BernH,


    I understand what you are pointing out with the Windows NTFS incompatibility, as well as the myriad of variables you pointed out with the various file system types.


    I guess my frustration is there isn't an easy "converter/adapter" either hardware or software, that can let you get to your data immediately.


    Raid / backup doesn't mean much if you have a handfull of heavy doorstops with all your data and you can't access it unless you have another full on system to read the drives, and know HOW to recover from failure.


    Also, YouTube is full of videos showing how to set up OMV/TrueNAS, but not how to deal with various failures, step by step, to walk noobs through the inevitable situation when the drive or hardware fails, then what?


    It's just a serious piece of the puzzle that's lacking.

    • Official Post

    #1. If your hardware fails (Motherboard, computer, power supply, whatever) and you can't find a replacement hardware, and you NEED access to your OMV data, can you take one of your OMV drives out and plug it into a Windows machine with a USB SATA adapter, can you copy your data off the OMV drive to your windows computer? (Yes I know Windows is NTFS)

    Generally speaking and by default, No. A popular / typical file format for a data drive, in Linux, is EXT4. With that said, there are a few EXT4 packages for Windows, like Ext2explore, that will allow Windows to read the EXT4 format. More -> info.

    #2. There isn't a straightforward guide that I can find that explains STEP BY STEP, if your OMV drive fails, the correct procedure to remove it, put a new one in, and OMV rebuilds the drive. (do you need to format the drive first? If so what format?)

    - Removing a drive, physically, is straight forward. Removing references to a failed drive is another matter.
    - Adding a new drive is covered here -> A Basic Data Drive. (You might want to at least skim through the -> New User Guide. It will answer at least some of your questions.
    - OMV does not "rebuild" drives. MDADM (software) RAID will rebuild a drive. Advanced file systems like ZFS or BTRFS or a package like SnapRAID will rebuild drives. All of these options are supported by OMV. There are general guides on the internet for MDADM RAID, ZFS and BTRFS drive restorations. Lastly, SnapRAID is an available as a plugin. The OMV-Extra's wiki has a SnapRAID restoration process in a -> SnapRAID doc. Of course, the forum supports questions on all of these topics.


    #3. If you want to expand storage space in the future, HOW is this done explained in a step by step method of removing the smaller drives and replacing with larger capacity drives.

    One at a time?

    Since expanding storage is a broad topic, which depends on how storage is set up at the beginning, there's no step-by-step document spelling it out. Software RAID5 can do what you're suggesting by failing a drive in the array (we'll say it's 4TB), replacing it with an 8TB drive, and rebuilding the array. Then that process is repeated for all array member drives. When all are replaced, the filesystem partition is expanded to make use of the extra space. For ease of drive expansion, mergerFS is a good choice that can be implement at any point in time, using drives that already have data on them. mergerFS will allow the addition of a drive, of any size, into it's pool at any time. We have a -> mergerFS doc as well.


    My very first initial experience with OMV is very disapointment, because trying to ready through the forums, I am seeing conflicting info, sarcasm, squables, attitide.

    While we try to keep it "G" rated, on any social platform some of this will happen. It's part of human nature, so it is what it is. Also note that while we have a lot of forum supporters, with a lot of knowledge, this forum is NOT about "Computer Science". It's about the "opinions" of experienced PC and server users, LAN admins, and Developers. There will be a number of varying opinions on how things should be done, in the Linux server world, just as there are differences of opinion on maintenance and repair techniques between different car mechanics.


    Just an FYI, I am not seeing much better on TrueNAS Scale for that matter.

    Not surprising.


    As my first OMV test, I just set up an HP T620 Thin client. I bought an M2 SSD to USB adapter and put a 1TB SSD in it, watched a few YouTube videos, and everything is working fine with it, however it just is a perfect single point of failure, and it was just as a test to see how OMV works and to get it on the network.

    This is what backup is for. While full data backup is possible using external drives with a drive dock, my personal preference is to have a full backup server. It's not as difficult or even as expensive as one might imagine. We have a doc for that as well.


    I don't think so. I have 3 full servers and 1 SBC with complete copies all data (two have ZFS mirrors, one has RAIDZ1, one has a BTRFS mirror). Another SBC has a subset of data. It's running MergerFS & SnapRAID. One of the servers is an old Intel server with dual Xeon's.

    Admittedly, some of that hardware is for test purposes (forum support) but I've had two servers since my beginning into Linux servers. I believe in backup.


    The CRAZY thing is... the simple single point of failure with the HP T-620 thin client and the USB M2 SSD...it's formatted as NTFS, which means I can grab that, plug it into a USB drive on any Windows machine, and have instant access to my data, which is my concern with question #1.

    There are real and potentially serious permissions issues when doing this. Linux (POSIX compliant permissions) do not, and can not map, one-to-one with Windows/NTFS permissions. If any Linux permission (other than Others - Read/Write/Execute - the rough NTFS equivalent of Everyone) is applied to NTFS files and folders, the potential for permissions issues exists. This applies to an OMV server or Truescale for that matter. It's true of any Linux server. For emergency data retrieval purposes, you've be better off using an EXT4 package with Windows for read access.

    On the other hand, I can't see 4 Linux servers failing all at once so using NTFS with data drives, simply because they can be plugged into Windows, seems like something of a stretch to me. (That's an opinion. :- )

    • Official Post

    jdredmond Welcome to the forum.


    They have answered most if not all of your questions. Although I would like to add the following:

    It sounds like your biggest concern is instant access to data if the OMV server fails. Regarding this, I just want to tell you that I would try to forget the idea of accessing data from Windows, there are multiple ways to regain access to data instantly.

    1. If you have another OMV server with a backup. In 5 minutes this backup server can replace the original server with access to the data.

    2. If you have another OMV server but the data is not backed up. You can mount the data hard drive on the second server and access the data instantly.

    3. If you don't have another server. It will take a little more time, but you can reinstall OMV on the failed server and remount the existing file system.

    4. If the above is not possible because the server is not working and you do not have a backup server. As a last resort you can start any Linux distro in test mode on any PC and access that data disk. This is relatively fast.


    The general idea is that once you have all your data on a server you will never need to access the data directly from a PC, you will always do it over the network. Depending on how you make your backups, access will be more or less quick in case of a failure. The general rule is to always have a complete backup of all data, there are several ways to achieve this.


    Once the importance of having a backup has been clarified, the issue of Raid or not Raid will depend on whether you really need it. In my case I use it for the following reasons:

    1. Availability. Some of the files on my server are work files and I can't afford to lose some newly created/modified files if a hard drive fails and the next scheduled backup hasn't happened yet.

    2. Protection against data corruption. I use ZFS to have that protection. mdadm does not protect against bitrot. It is probably very difficult to have problems due to bitrot, but since I configure a Raid I prefer a file system that also provides me with that protection.

    Another reason someone might argue for setting up a Raid is faster pool access speed, although if you have a gigabit network this probably makes no sense.


    On the other hand, as you have seen in the answers that have already been given, there are infinite ways to configure a server, file systems and backups. So it is difficult to generalize a detailed procedure that works for everyone. Don't forget that behind OMV is Debian and the options it offers are many.

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!