Upgrade from OMV 4 to 7 / RAID1 VS RAID5 / Data Migration?

    • New

    Hi OMV community,


    I have installed Open Media Vault (OMV) 4 on a HP MicroServer Gen8 back in 2018.

    It contains a 2.5'' SATA SSD for the Open Media Vault (OMV) 4 operating system and 2x 4TB 3.5'' SATA HDDs in RAID1 for backup.


    Every couple of days/weeks I power up the HP MicroServer Gen8 via Wake On LAN (WOL),

    then initiate a backup by automatically copying the new/updated files from my PC to it over LAN using a pre-defined FreeFileSync batch job.


    So far it worked without any problems,

    but now I'm running out of storage space and would need some general guidance how to proceed.



    1.) Upgrade from OMV 4 to 7:


    Is it possible/recommended to perform an upgrade from OMV 4 to the latest version OMV 7?


    If yes, I would follow the steps included within the following thread (RE: Is there a guide to in place upgrade?) :

    (Note: In addition, after step 1.) I would also try to update the BIOS and firmware on the HP MicroServer Gen8 accordingly.)



    2.) RAID1 VS RAID5:


    As mentioned above, currently OMV which is installed on a HP MicroServer Gen8 contains 2x 4TB 3.5'' SATA HDDs running in RAID1 for backup but I'm running out of storage space.


    The HP MicroServer Gen8 only has 4x 3.5'' drive bays in total.

    So my plan is to replace these 2x 4TB 3.5'' SATA HDDs running in RAID1,

    with 4x new 3TB 3.5'' SATA HDDs running in RAID5 instead.


    But is switching from RAID1 to RAID5 a bad idea for home user backup (emails, photos,...) in terms of reliability and data recovery?



    3.) Data Migration:


    As the HP MicroServer Gen8 only has 4x 3.5'' drive bays in total,

    how do I migrate the data from the existing 2x 4TB 3.5'' SATA HDDs running in RAID1 to the new 4x 3TB 3.5'' SATA HDDs running in RAID5?


    If I remember correctly, the old 2x 4TB 3.5'' SATA HDDs running in RAID1 are using NTFS as a file system.


    For example is it possible to just plug one of the old 4TB 3.5'' SATA HDDs into my PC,

    and then copy the data over LAN via FreeFileSync to the new 4x 3TB 3.5'' SATA HDDs running in RAID5 on the HP MicroServer Gen8?




    Thanks and best regards

  • XYZ1

    Added the Label OMV 4.x
    • New

    IMHO The easiest thing would be to get an external drive that is big enough to take all your data. Then backup the data to that. Set up a new server in the configuration you want and then copy the data to that new drives.


    Upgrading etc with a life system is possible, but also has a significant chance to produce challenges.

    Quote from XYZ1

    running in RAID1 for backup

    Raid is no backup. Raid is only for reliability and/or speed. If you do not need that, use a lvm for maximum space. If you need a backup, then make a proper one one a dedicated drive.

    • New

    Quite a lot to think about.


    There's no direct upgrade path from OMV4 to OMV7. TBH, if your OMV config is relatively simple, I'd just create a new OMV install using latest OMV7 and then configure that.


    If you move from using 2 to 4 HDD bays, you need to think about how you will boot your microserver and where you install OMV. (For example ssd in the optical drive bay, booting form a separate USB mem stick which holds the GRUB bootloader, or using the internal USB header with a USB to SATA cable to connect to your SSD, or actually running from OMV from a USB stick or SD card. )


    How much storage do you need? Do you already have the 4 x 3TB drives? There may be better 3 drive combos that allow to keep your SSD in the leftmost bay and install on and boor from your SSD.


    Data migration will depend to some extent what drive config you settle on and will probably need some juggling. But the big question here is the continued use of FreeFileSync. Correct me if I'm wrong, but I thought this software was definitely windows file system to windows file system. So are you mounting an OMV SMB share to your PC before you execute your batch job? OMV can only make very limited use of NTFS.

    • New

    First of all, thank you very much for your replies!


    Quote from Steini

    IMHO The easiest thing would be to get an external drive that is big enough to take all your data. Then backup the data to that. Set up a new server in the configuration you want and then copy the data to that new drives.

    I had a look and according to the OMV 4 GUI ("Storage" -> "File Systems") the old 2x 4TB 3.5'' SATA HDDs running in RAID1 are not using NTFS but ext4 as the file system.


    My Windows 11 based PC is not able to read ext4 formatted HDDs by default.

    So for data migration it would not be possible to just plug one of the old 4TB 3.5'' SATA HDDs into my PC,

    and then copy the data over LAN via FreeFileSync to the new 4x 3TB 3.5'' SATA HDDs running in RAID5 on the HP MicroServer Gen8.

    This would require some kind of additional error-prone steps (using WSL,... (https://windowsreport.com/ext4-windows-11/)).


    But possible I want to avoid buying an external >=4TB USB HDD for this kind of one time data migration step.

    I also have an external 3.5'' SATA to USB enclosure.

    So would it be possible to just plug one of the old 4TB 3.5'' SATA HDDs into it,

    directly connect it to the HP MicroServer Gen8 via one of the USB ports and then copy the data to the new 4x 3TB 3.5'' SATA HDDs running in RAID5?



    Quote from Krisbee

    There's no direct upgrade path from OMV4 to OMV7. TBH, if your OMV config is relatively simple, I'd just create a new OMV install using latest OMV7 and then configure that.

    Okay, so in case I try to perform the upgrade from OMV 4 to 7 instead of doing a re-installation of OMV 7 from scratch,

    I need to follow the steps mentioned above after every single upgrade (OMV 4 -> 5 -> 6 -> 7) as it would not be possible to skip versions.


    Maybe I create a backup of the 2.5'' SATA SSD containing the OMV 4 operating system first and try to perform the upgrade.

    In case it fails I would do a re-installation of OMV 7 from scratch.



    Quote from Krisbee

    If you move from using 2 to 4 HDD bays, you need to think about how you will boot your microserver and where you install OMV. (For example ssd in the optical drive bay, booting form a separate USB mem stick which holds the GRUB bootloader, or using the internal USB header with a USB to SATA cable to connect to your SSD, or actually running from OMV from a USB stick or SD card. )

    This would not be an issue, as the 2.5'' SATA SSD containing the OMV 4 operating system is mounted within the ODD drive bay and I'm also using a SD card containing the GRUB bootloader already.



    Quote from Krisbee

    How much storage do you need? Do you already have the 4 x 3TB drives? There may be better 3 drive combos that allow to keep your SSD in the leftmost bay and install on and boor from your SSD.

    As with the current 2x 4TB 3.5'' SATA HDDs in RAID1 I'm running out of storage space,

    I would at least need 6TB of storage space.

    Yes, I already have the 4x new 3TB 3.5'' SATA HDDs.



    Quote from Krisbee

    Data migration will depend to some extent what drive config you settle on and will probably need some juggling. But the big question here is the continued use of FreeFileSync. Correct me if I'm wrong, but I thought this software was definitely windows file system to windows file system. So are you mounting an OMV SMB share to your PC before you execute your batch job? OMV can only make very limited use of NTFS.

    Yes, I had a look and according to the OMV 4 GUI ("Storage" -> "File Systems") the old 2x 4TB 3.5'' SATA HDDs running in RAID1 are not using NTFS but ext4 as the file system.
    Every couple of days/weeks I power up the HP MicroServer Gen8 via Wake On LAN (WOL) which is providing a SMB share to my Windows 11 PC,

    then initiate a backup by automatically copying the new/updated files from my PC to it over LAN using a pre-defined FreeFileSync batch job.




    Thanks and best regards

    • New

    XYZ1 What a pity you haven't got new 4TB's rather than new 3TB HDDS, as with two free bays in you microserver migrating the data could have relatively straightforward. I'm I right in assuming that more than 3TB of data is stored on OMV.


    You can't simply use one of the old 4TB while its part of a RAID1. It is possible to remove a single drive from a RAID1 mirror and re-format it in OMV but it leaves the RAID1 running in degraded mode and your data is at risk. So how precious is the data stored on OMV? Is losing it all just an inconvenience, or a disaster?

    • New
    Quote from Krisbee

    XYZ1 What a pity you haven't got new 4TB's rather than new 3TB HDDS, as with two free bays in you microserver migrating the data could have relatively straightforward. I'm I right in assuming that more than 3TB of data is stored on OMV.

    Yes, I'm running out of storage space on the existing 2x 4TB 3.5'' SATA HDDs in RAID1.


    But I got the 4x new 3TB 3.5'' SATA HDDs relatively cheap and thought it might be a good option to run them in RAID5.


    But how do you mean with 2x new 4TB 3.5'' SATA HDDs instead, data migration could have been relatively straight forward?:

    - Add one of the new 4TB 3.5'' SATA HDDs into the HP MicroServer Gen8

    - Temporarily copy the data from the old 2x 4TB 3.5'' SATA HDDs running in RAID1 onto the new 4TB 3.5'' SATA HDD

    - Create a new RAID5 with 3x 4TB 3.5'' SATA HDDs (=based on 2x old ones and 1x new one)

    - Copy the data temporarily stored on one of the new 4TB 3.5'' SATA HDD to the new RAID5 with 3x 4TB 3.5'' SATA HDDs



    Quote from Krisbee

    You can't simply use one of the old 4TB while its part of a RAID1. It is possible to remove a single drive from a RAID1 mirror and re-format it in OMV but it leaves the RAID1 running in degraded mode and your data is at risk. So how precious is the data stored on OMV? Is losing it all just an inconvenience, or a disaster?

    Okay, but how is the handling of a RAID1 array supposed to be in case one the HDDs dies?

    My assumption was that in this case you can either rebuild the RAID1 by swapping the defective HDD.

    Or simply take the other HDD which ist still working and backup the data to another safe location.

    • New

    Yes, something along those lines assuming you stick to MD RAID, there are other options for disk pooling in OMV. The disk juggling gets more involved using 3TB HDDs as the target.


    This comes back to what Steini said about RAID not being a backup. One disk dies and RAID1 becomes degraded, a further failure and your data is lost. Data is still available in degraded mode but you don't really want to run this way for a prolonged period without redundancy. Using the working disk elsewhere needs another working Linux system with the MD RAID software installed.


    For your current use case, you could argue using one 4TB in the NAS and another in a 4TB USB enclosure attached to the OMV when needed was a good option. Backup from PC to OMV and then from OMV internal drive to external USB drive attached to OMV, but not necessarily at the same frequency..

    • New

    Okay so at the end RAID1 doesn't really seem to work the way I originally thought.

    Which means in case one of the HDDs of a RAID1 array dies,

    you have to swap the defective HDD (to ideally the same model again) and hope that it would be possible to successfully rebuild the RAID1 array.

    As an alternative, it is not possible/safe to simply take the other HDD which is still working and backup the data to another safe location.


    But this also means (to answer my own question from above) it would be okay to switch from RAID1 to RAID5,

    as at the end "RAID is not a backup" and in addition to my NAS I would need to have another backup somewhere else anyway.


    I'm still facing the challenge how to do the data migration of the old 2x 4TB 3.5'' SATA HDDs running in RAID1 but I will find a solution.

    Once I have temporarily copied the data from the old 2x 4TB 3.5'' SATA HDDs running in RAID1,

    I will try to update the BIOS and firmware on the HP MicroServer Gen8 accordingly.

    Afterwards I will install OMV 7 from scratch on the HP MicroServer Gen8 and create a RAID5 array based on the 4x new 3TB 3.5'' SATA HDDs.

    Once this RAID5 array has successfully created I will copy the old data back to it.


    As mentioned above, in addition to my NAS I would need to have another backup somewhere else also.

    • New

    @XYZI While your thinking about data migration, one possibility is to temporarily turn your PC into a NAS by:


    1. Installing the latest OMV on a USB stick

    2. Disconnecting drives on the PC after shutting down.

    3. With the HP gen8 shutdown, remove the two 4TB drives and connect them to spare/unused SATA ports on your PC.

    4. Install OMV7 from scratch on the HP MicroServer Gen8 and create a RAID5 array based on the 4x new 3TB 3.5'' SATA HDDs, etc.

    5. Boot OMV on your PC, check the RAID1 is detected and its filesystem can be mounted. Use "rsync" to copy data from RAID1 to your new RAID5.

    6. On completion, detach the two 4TB drives from your PC and return it to normal use.


    A useful rsync ref: https://noted.lol/how-to-sched…make-copies-of-your-data/


    If you preferred to do everything on the microserver, then you could:


    1. Using the current OMV4, install 2 x 3 TB drives. Create and mount an EXT4 filesystem on each drive.

    2. Copy the 4TB of data held on existing RAID1 , split across the two 3TB drives, say 3TB + 1TB.

    3. Via OMV remove one 4TB drive from the RAID1, putting it into degraded mode.

    4. Secure Wipe the single 4TB drive you removed from the RAID1 and create an EXT4 filesystem on this drive.


    At this point you have backed up the RAID1 data to two 3TB drives and freed a single 4TB drive.


    5. Remove the 4TB drive that is still part of the degraded RAID a place itaside for safekeeping.

    6. Copy the 4TB worth of data back from the two 3TB drives to the single 4TB drive that remains in the microserver.

    7. Shutdown and remove all 3 drives from the microserver, be careful to keep the 4TB separate.

    8. Install OMV7 to the microserver with no hard drives in situ. ( update, set ip , and "user" account, test ssh as necessary)

    9. Shutdown and then put the single 4TB drive ( not the RAID1 disk) and three 3TB drives in the microserver.

    10. Boot OMV7. Wipe all 3TB drives. Create a 3 x 3TB RAID5, and create EXT4 filesystem on the RAID5. Copy the data on 4TB drive to RAID5. Unmount 4TB drive. Shutdown OMV7.

    11. Replace 4TB drive with fourth 3TB drive. Boot OMV7. Wipe 3TB drive (make sure you select the drive just added to server). Grow RAID5 by adding the 4th 3TB drive. Grow RAID filesystem.


    End result is 4 x 3TB RAID5 of 9TB capacity with 4TB used, a RAID5 can sustain the loss of only a single drive. You will additionally have one 4TB (not part of a RAID) with a copy of the original data and a second 4TB ( the degraded part of the RAID1) with a second copy of the original data.


    Doing everything on the microsever is time consuming and prone to error unless you keep very careful track of which drives are which in the server. Names like /dev/sda, /dev/sdb etc can point to different disks as they are swapped in an out. Go by the "serial numbers" shown in the OMV WebUI "disk" section. Make a careful note of the data on the physical drive labels before you start.

    • New

    XYZ1


    The data migration become simpler if you are prepared to used a BRTFS RAID1 profile instead of MD RAID5. This would mean starting by installing OMV7 on your server when all the HDDs were removed. In outline, the procedure to then follow is:


    1. Put the two 4TB RAID1 back, along with two 3TB HDDs. Check the RAID1 is recognised and the filesystem it holds mounts.

    2. As above in #9, create an EXT4 on each individual 3TB HDD and copy the 4TB of data held on existing RAID1, split across the two 3TB drives, 3TB + 1TB.

    3. Shutdown and remove the two 4TB drives and add the two remaining 3TB drives.

    4. Boot OMV7, creates a BRTFS RAID1 on the two 3TB drives just added. Copy the 3TB of data to this new BRTFS RAID1.

    5. The 3TB HDD EXT4 drive's filesystem can now be unmounted and the drive wiped.

    6. The drive just wiped can be now added to the BTRFS RAID1 via a CLi command, which increases its capacity to 4.5TB.

    7. Copy the remaining 1TB of data to the now 3x 4TB BTRFS RAID1.

    8. With all the original 4TB of data now on the 3 x 3TB BRTFS RAID1, unmount the 3TB drive that held the 1TB of data and wipe it.

    9. Finally, the drive just wiped can be now added to the BTRFS RAID1 via a CLI command, creating a 4 x 3 TB BTRFS RAID1 with a capacity of 6TB.

    • New

    Thank you very much for these detailed instructions how data migration could be done!


    The 3rd option you mentioned with BRTFS RAID1 seems to be quite comfortable.

    However, I have read some mixed opinions about BRTFS and in the end it would only be 4x 3TB BTRFS RAID1 with a capacity of 6TB, instead of 4x 3TB RAID5 with a capacity of 9TB.

    So I would probably have to repeat more or less the same feat of data migration in the next few years, as file sizes are increasing more and more (e.g. videos recorded with a smartphone in 4K@60Hz,...).


    So I think I will go for the following option:

    • New

    XYZ1 The chosen option has the distinct advantage of keeping the existing RAID1 intact and can you prep your microserver with OMV7 within your normal two week window for what should be one-off data migration.


    As your only route to future capacity is expansion is by using bigger drives, I wouldn't dismiss BRTFS too quickly as it does allow you use drives of mixed sizes. For example, https://carfax.org.uk/btrfs-us…4000&d=4000&d=3000&d=3000 So unless your 4TB power on hours make the candidates for replacement, once the data migration is complete you could actually swap disks around in a BTRFS RAID1 to get a capacity of 7TB.


    Personally, I'd want to move away from MD RAID5 ( raid5 write hole and lack of data integrity) another you could consider using ZFS on OMv7. A four disk RAIDZ1 gives you 9TB capacity with single disk redundancy, a RAIDDZ2 gives you 6TB capacity with two disk redundancy. A four disks arranged in the equivalent of RAID10 gives you 6TB capacity with two disk redundancy as long as you don't lose both sides of a single mirror.

    • New
    Quote from Krisbee

    As your only route to future capacity is expansion is by using bigger drives, I wouldn't dismiss BRTFS too quickly as it does allow you use drives of mixed sizes. For example, https://carfax.org.uk/btrfs-us…4000&d=4000&d=3000&d=3000 So unless your 4TB power on hours make the candidates for replacement, once the data migration is complete you could actually swap disks around in a BTRFS RAID1 to get a capacity of 7TB.

    Okay, I was not aware that a RAID1 can be done with 4 disks if using the file system BTRFS instead of EXT4!
    I was always thinking a RAID1 can only be done with 2 disks in general.


    But is my understanding correct that the reason why such a RAID1 can be done with 4 disks (if using BTRFS RAID1 instead of EXT4 RAID1),
    is that it's not true disk mirroring anymore but instead it is based on block-level redundancy (https://www.reddit.com/r/btrfs…do_you_really_need_raid1/ )?:

    "Btrfs RAID-1 isn't true mirroring. It's block-level redundancy, not device-level. While not a new or unique idea, it's still pretty cool - and in many important ways, a big improvement over device-level mirroring."


    If yes as it is not true disk mirroring anymore but instead it is based on block-level redundancy only,
    the amount of redundancy is not as good anymore, right?


    I have also seen some recommendations to not use BTFRS in RAID:

    Modernize your Linux Storage with btrfs!

    External Content www.youtube.com
    Content embedded from external sources will not be displayed without your consent.
    Through the activation of external content, you agree that personal data may be transferred to third party platforms. We have provided more information on this in our privacy policy.




    Quote from Krisbee

    Personally, I'd want to move away from MD RAID5 ( raid5 write hole and lack of data integrity) another you could consider using ZFS on OMv7. A four disk RAIDZ1 gives you 9TB capacity with single disk redundancy, a RAIDDZ2 gives you 6TB capacity with two disk redundancy. A four disks arranged in the equivalent of RAID10 gives you 6TB capacity with two disk redundancy as long as you don't lose both sides of a single mirror.

    So to summarize RAIDZ1 would definitely be the better option than RAID5, right?


    I have read that compared to TrueNAS the ZFS file system is not included within OMV by default and must be added via the optional plugin OMV Extras (How do I get ZFS in OMV?).
    So is it save/reliable to use the ZFS file system in OMV, or is it more like experimental?


    I have also read that the ZFS file system heavily depends on the amount of available RAM.

    The HP MicroServer Gen8 only contains a Intel(R) Celeron(R) CPU G1610T (=2.30GHz / 2 cores / 2 threads) and 2x2GB ECC RAM,

    so would that be sufficient for the ZFS file system / RAIDZ1 at all?

    • New

    XYZ1


    To answer your questions:


    The redundancy level of a BTRFS RAID1 profile is the same as a MD RAID5. Its survives the loss of one disk only. MD RAID5 uses trad disk based stripes with parity, BTRFS RAID1 uses a space allocation based on "chunks" which are then filled with "extents". As you have read, it's the "chunks" which are "mirrored" on two different drives out the N drives in BTRFS RAID1, where N is >=2.


    The unqualified statement re: BRTFS RAID in the vid is nonsense. BTRFS RAID5/6 is not production ready, but do you really think OMV's founder would make BTRFS available in OMV if it were unsafe? Looking at the number of people who post on this forum with MD RAID5 woes, BTRFS could have been a better choice.


    ZFS in OMV is most definitely not experimental. In OMV7 MD RAID has also been moved to a plugin. It is true that ZFS makes extensive use of a special RAM cache feature called ARC (Adaptive Replacement Cache) and the size of the ARC can impact performance but that very much depends on the data access pattern on your pool - e.g. the number concurrent read/write requests, random versus sequential access, etc. and the level of performance needed. For your use case of a basic file server, streaming writes to disk with occasional reads all over what I assume to be a 1Gbe network, your microserver specs are enough.


    I don't know how your data on OMV is currently organised. Perhaps it is just a single shared folder that is mounted in Windows which contains all the directories needed. Clearly, the MD RAID5 path is the most straightforward to adopt. But you do have the chance to try out/use BTRFS or ZFS. In which case, there an opportunity to review how your data is organised in terms of shared folders and network shares. In BTRFS, each new "shared folder" is created as a "subvolume". In ZFS, you'd create one or more "filesystem datasets" on your zfs pool and link each dataset to a "shared folder".

    • New

    I decided to proceed as follows and create a BRTFS RAID1:


    I managed to perform step 1., 2. and 3.,
    but regarding the step "4. Boot OMV7, creates a BRTFS RAID1 on the two 3TB drives just added...." I'm not sure if I have created the BRTFS RAID1 correctly?:

    - Verified if the 2x new 3TB 3.5'' SATA HDDs which I just added are visible within OMV:

    "Storage" -> "Disks"
    -> Okay, the 2x new 3TB 3.5'' SATA HDDs are visible.


    - Formatted the 2x new 3TB 3.5'' SATA HDDs as follows within OMV:

    "Storage" -> "File Systems" -> "Create and mount a file system" -> "BTFRS":

    - "Type": "BTRFS"

    - "Profile": "RAID1"

    - "Devices": 2x new 3TB 3.5'' SATA HDDs selected


    Initially I tried to create this BRTFS RAID1 via "Storage" -> "Multiple Device" but could not find the respective entries for BRTFS RAID1 at all.
    Therefore I tried to create this BRTFS RAID1 via "Storage" -> "File Systems" -> "Create and mount a file system" as mentioned above.
    Based on my understanding most likely the reason why such a BRTFS RAID1 must be created via "Storage" -> "File Systems" -> "Create and mount a file system" and not via "Storage" -> "Multiple Device" is, that it's not a kind of classic RAID anymore.


    However, I'm still confused if I have created this BRTFS RAID1 via "Storage" -> "File Systems" -> "Create and mount a file system" correctly,
    as "Storage" -> "File Systems" now shows the following information:

    Gekjsit.jpeg


    As this this BRTFS RAID1 currently is only based on 2x new 3TB 3.5'' SATA HDDs I was expecting to see a available capacity of ~2-3TiB but not 5.46TiB?

    • New

    XYZ1


    Via ""Storage" -> "File Systems" -> "Create and mount a file system" is the correct way.


    Highlight the BTRFS filesystem as in your screenshot and then click on the triangle symbol to see the details of the BTRFS RAiD1. Adjusting the columns displayed and tagging the filesystem can give a more meaningful list. For example:



    The available storage figure will make more sense when you've written some data to the the filesystem.

    • New

    Thanks for the confirmation that I have created this BRTFS RAID1 correctly! :)


    In the meantime I also did the remaining part of step 4. "...Copy the 3TB of data to this new BRTFS RAID1.",
    by creating a new shared folder on the BRTFS RAID1 and copying the files via a local rsync task to it.
    As you mentioned, the available storage figure makes more sense now once I've written some data to the filesystem!



    Regarding step 6. "The drive just wiped can be now added to the BTRFS RAID1 via a CLi command, which increases its capacity to 4.5TB.":


    I can see the newly created BTRFS RAID1 (=based on 2x new 3TB 3.5'' SATA HDDs) and the remaining 2x new 3TB 3.5'' SATA HDDs which are used to temporarily store the data within OMV via "Storage" -> "File Systems" now as follows:
    loKcsyx.jpeg

    (Note: The data has been successfully copied from the 1st (=/dev/sde1) of the remaining 2x new 3TB 3.5'' SATA HDDs to the newly created BTRFS RAID1 (=/dev/sdb).)


    So is my understanding correct that to add the 3TB 3.5'' SATA HDD to the BTRFS RAID1 I would need to proceed as follows?:


    - Access the CLI of OMV via SSH and log in as root


    - Enter the following command to identify the mount point of the BTRFS RAID1:

    sudo lsblk

    (Note: In my case the mount point of the BTRFS RAID1 would be as follows, right?:
    "/srv/dev-disk-by-uuid-73180add-31d0-4cac-b6ce-8a21 8144cd63")


    - Wipe the 3TB 3.5'' SATA HDD within OMV via "Storage" -> "Disks" -> select "/dev/sde" -> press the "Wipe" button


    - Enter the following command (sudo btrfs device add <device path> <mount point>):

    sudo btrfs device add /dev/sde1 /srv/dev-disk-by-uuid-73180add-31d0-4cac-b6ce-8a21 8144cd63


    - Enter the following command:

    sudo btrfs filesystem balance /srv/dev-disk-by-uuid-73180add-31d0-4cac-b6ce-8a21 8144cd63 &

    • New

    XYZ1


    The BTRFS mount point you quoted is wrong, you have too many digits and there would be no space. It looks to be:

     /srv/dev-disk-by-uuid-73180add-31d0-4cac-b6ce-8a21 but please double check.


    You can get the long UUID part from the BTRFS filesystem details given by the WEBUI. Adding the "/srv/dev-disk-by-uuid-" prefix to the UUID gives the actual mount point.


    CLI commands that give the mount point include cat /etc/fstab and findmnt --real

    • New

    First of all thank you very much for supporting me during this journey!



    1.) In the meantime I've managed to perform the following steps:


    - Update BIOS and firmware on the HP MicroServer Gen8

    - Install OMV 7.X on the HP MicroServer Gen8

    - Add the new 4x 3TB 3.5'' SATA HDDs to the HP MicroServer Gen8 and create a BTRFS RAID1

    - Copy the data from the old 2x 4TB 3.5'' SATA HDDs EXT4 RAID1 to the new 4x 3TB 3.5'' SATA HDDs BTRFS RAID1 via Rsync

    - Add the required shared folders, users and SMB shares

    - Use WOL to wakeup the the HP MicroServer Gen8

    - Use FreeFileSync to backup the data from PC to the HP MicroServer Gen8 (=4x 3TB 3.5'' SATA HDDs BTRFS RAID1)


    - Plug-in old 2x 4TB 3.5'' SATA HDDs in EXT4 RAID1 into an external USB HDD enclosure and connect it to one of the USB 3.0 ports of the HP MicroServer Gen8

    - Mount existing old 2x 4TB 3.5'' SATA HDDs in EXT4 RAID1 file system

    - Create a local Rsync task within OMV to initiate a backup from new 4x 3TB 3.5'' SATA HDDs BTRFS RAID1 to old 2x 4TB 3.5'' SATA HDDs


    - In addition I'm thinking about using Microsoft 365 to upload at least the most important data (encrypted) to Microsoft OneDrive



    2.) According to the OMV GUI (="Storage" -> "File Systems" -> Button "Show details" the 4x 3TB 3.5'' SATA HDDs BTRFS RAID1 looks as follows:


    Does this look okay?

    Would it make sense to access the CLI of OMV via SSH, log in as root and enter the following command again/periodically?:

    sudo btrfs filesystem balance /srv/dev-disk-by-uuid-73180add-31d0-4cac-b6ce-8a21 8144cd63 &



    3.) I have read something about BTRFS RAID1 VS. BTRFS RAID1C3 (https://www.reddit.com/r/btrfs…aid1c3_metadata_question/),

    whereas BTRFS RAID1C3 seems to be more stable in terms of possible data loss.

    So would it make sense to use BTRFS RAID1C3 instead of BTRFS RAID1?

    • New

    XYZ1


    Your BTRFS filesystem look OK. The "details" concatenates output from three separate BTRFS commands:


    btrfs fi sh mountpoint

    btrfs fi df mountpoint

    btrfs dev stats mountpoint


    The output of the last command should show dev stats for all four individual devices, but is curtailed. If you want a fully picture of the way data and metadata have been allocated. etc use: btrfs fi us mountpoint. For example:


    Details from the WebUI;



    Output of btrfs fi us command:



    As you built and copied data to your 4 device BRTFS RAID1 filesystem ins stages, I'd do one full balance now. It will take a couple of hours to complete. You should not need a full balance in the future unless disks are replaced. A regular "filtered" balance of data only at say 5%, in step with your backup schedule would be useful. See: https://wiki.tnonline.net/w/Btrfs/Balance.


    Stick with RAID1 for data and metadata.


    Running into a "no space left" error (ENOSPC) on BTRFS can be a pita. A kind of catch-22, when you can't delete stuff because there's not enough unallocated space to increase the number of metadata chunks needed for the deletions. A consequence of BTRFS first allocating space in 1GB chunks, before filling those with extents. You can be bitten by ENOSPC during a balance too, because of the "working space" it needs. See: https://wiki.tnonline.net/w/Btrfs/ENOSPC


    So, enable email notifications and "edit" your filesystem in the WEBUI to set a usage warning threshold at say 85%.

Participate now!

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

Register Yourself Login