OMV 3.X versus OMV 2.X; "Growing" a RAID 5 array

    • Offizieller Beitrag

    (Per a forum user conversation in another thread) Grow RAID 5


    After testing RAID 5 scenarios in VM's of OMV 3.0.73, versus a real world platform running OMV 2.1:
    _______________________________________________________________________


    It seems that the behavior of the "Grow" button in <System>, <RAID Management> has changed between OMV versions 2.X and 3.X.


    Where the "Grow" button in OMV 2.1 triggers an array restrip operation that integrates a new drive, the "Grow" button in 3.X versions simply adds a drive to the array as a hot spare which is more in line with the mdadm --add command line.


    To integrate a disk into an OMV 3.x array (Grow it) it was necessary to do it from the command line (as follows).
    mdadm --grow /dev/md0 --raid-devices=4


    Perhaps there's a reason for the change?
    ______________________________________________________________________


    If restriping (or Growing) a RAID 5 array is to be supported in OMV 3.X , it makes sense to attach the CLI operation (mdadm --grow /dev/md0 --raid-devices=4) to the Grow button in RAID Management.
    Similarly, when adding a hot spare to an array, it would be more intuitive if a new GUI "Add" button is attached the CLI operation (mdadm --add /dev/md0 /dev/sde).


    Are we off base or are we missing something? In any case, any information or an explanation would be appreciated.



    Thanks

    • Offizieller Beitrag

    It's still outstanding but I've come to understand why. When it comes to growing an mdadm RAID array, users tend to "click" without contemplating the consequences and get in over their heads. Then, with a dead or dying array (usually without backup), they return to the forum after compounding the problem. Often, it's unsolvable.


    If you want to grow your array, you can find the command lines to do it here. - > mdadm RAID 5 - 3 disks to 4 disks


    If something goes wrong (that's not critical), here's a resource to recover your array.
    Recover array (There's good general reading at this site.)


    _______________________________________________________


    With the above noted - if you don't have a full backup of all the contents on your array, or at least the data that you can't afford to lose, don't try to grow it. Arrays die in these processes all the time. You can search this forum and others for more than a few examples.

  • Thanks for taking the time to reply, appreciated.


    Yes I've already manually got my RAID array reshaping in the background (additional 6Gb disk = couple of days likely!), just wanted to see whether or not this was a bug as, like other posters, I've done this several times before all in the GUI. Makes sense what you said and no drama to revert to the command line - fingers crossed my additional storage will satisfy me for a while (as additional storage at this sort of size starts to get expensive!).


    Thanks again.

    • Offizieller Beitrag

    Here's to hoping nothing goes south during the reshape. (It's not over until it's over.) If it comes out alright, as a last step you'll have to go into Storage, File Systems and "Resize" your file system to use the additional space.
    _______________________________________


    If adding a 6TB disk, I'm guessing you have a sizeable array. In any case, give some thought to backup even if that would mean backing up individual shared folders to an external drive or host. (They're making reasonably priced 8TB externals these days.)


    Keep in mind - the more disks you put into the array AND, as the existing drives grow older, the probabilities of a catastrophic failure increase. Just a thought, not a sermon. :)

    • Offizieller Beitrag

    I could never find any instruction on how to grow my RAID5 via GUI.

    Did you search the forum? Normally, adding the disk and clicking grow in the raid tab and resize in the filesystems tab is all that is needed.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | 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!

  • I did it manually several times in the console. Clicking grow did not add the disk. I had this issue the last time, also, I think. This time, I first put a filesystem on the disk and shared it via samba to check the disk (it had SMART status red and I wanted to re-allocate all broken sectors fist - this worked, SMART is green now, no pending sectors and no errors.)


    So, mayby I did not free this disk enough to make it addable to the RAID. But I did try to unmount it and delete the filesystem. But every time I want to add a disk I have troubles with it in the GUI because it is not that clear what to do. Or, when clicking on "grow" and nothing happens, there is no indication if this step should be possible or what could be the issue.


    EDIT: I searched google which brings up the forum a lot of times.

    • Offizieller Beitrag

    Or, when clicking on "grow" and nothing happens, there is no indication if this step should be possible or what could be the issue.

    If you are putting a filesystem on the disk first, that is a problem. You should be wiping the disk if there is anything on it. Most new disks come clean so this step isn't necessary. And when you grow an array, a dialog will pop to ask you which disk(s) to add. if you didn't see this, you didn't have disks available (probably because it had a filesystem on it) and it definitely did not grow your array. While I think the documentation could be better, I really don't think most users need raid.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | 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 clarification. Yes, most of my HDDs are used. What kind of "wiping" would be sufficient? Because I also tried to unmount the disk, delete the filesystem and then still the disk was not available to the RAID. Maybe a restart would have made a difference?


    Also, thanks for talking about RAID. This is a bit off topic, but I want to explain why a RAID is a good thing in my eyes: I got me openmediavault because I was planning to have a RAID. I am using my RAID6 (I misspelled it earlier) without any backup. This is just for movies. Some of them I still have on DVD/Bluray. So, my RAID gives me a first wall of security against hardware failure. Backing everything up would cost > 300 Euros at the moment. Bying the movies again would maybe cost tripple the amount. But regarding the error rate it is probably better without backup over all from a cost perspective. Since the backup in case of movies is on all the circulating dvds/bluerays.

    • Offizieller Beitrag

    What kind of "wiping" would be sufficient?

    Quick wipe from the physical devices tab should be fine.


    Because I also tried to unmount the disk, delete the filesystem and then still the disk was not available to the RAID. Maybe a restart would have made a difference?

    Nope. It needs to be wiped.


    Also, thanks for talking about RAID. This is a bit off topic, but I want to explain why a RAID is a good thing in my eyes: I got me openmediavault because I was planning to have a RAID. I am using my RAID6 (I misspelled it earlier) without any backup. This is just for movies. Some of them I still have on DVD/Bluray. So, my RAID gives me a first wall of security against hardware failure. Backing everything up would cost > 300 Euros at the moment. Bying the movies again would maybe cost tripple the amount. But regarding the error rate it is probably better without backup over all from a cost perspective. Since the backup in case of movies is on all the circulating dvds/bluerays.

    I don't think raid is giving you what you think you are getting. yes, it is protecting you from a failed drive but not from an accidental delete or corruption or crypto virus, etc. Personally, I think you would be better off to use unionfilesystems to pool the disks and snapraid with one or two parity disks. Either way, backup should be a priority instead of extra parity disks that are only giving you redundancy.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | 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

    Is there a way to transfer easily from a RAID to SnapRaid? Can you recommend a guide?

    Nope. You need to wipe your disks and format them with a filesystem, mount them in the filesystems tab, and then restore your data from backup. There is no guide for this because there are too many unique setups to try and account for.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | 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!

  • I'm having issues getting my RAID to grow and I'm pretty sure I did it the way I always have in the past except I didn't turn my machine off to do it.


    Inserted new 4tb red disk. Wiped it, selected grow under raid and added it to the raid. It still shows up clean but it never reshaped or anything, it shows as a spare under details, capacity never changed either. and growing file system hasn't done anything.


    Should I reboot? I'm on OMV3.


    Thanks!

    • Offizieller Beitrag

    First, do a full backup because any operation that changes an array can end in disaster


    (Earlier in this thread, command lines to do this are provided.)


    mdadm --grow /dev/md0 --raid-devices=4

    (Where md0 is the device name assigned to the array, and an existing spare is being added as a 4th drive in the array.)



  • I wasn't sure about the doing it via command line because @ryecoaaron was saying it shouldn't be necessary and the GUI should work.


    This is my 7th drive. I should change the 4 in the command to 7 right? Is it okay that it's already added via the GUI?

    • Offizieller Beitrag

    shouldn't be necessary and the GUI should work.

    There are "should"s in the statement :) While it should work, it may do something wrong.

    omv 7.0-32 sandworm | 64 bit | 6.5 proxmox kernel

    plugins :: omvextrasorg 7.0 | kvm 7.0.9 | compose 7.0.9 | 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

    This is my 7th drive. I should change the 4 in the command to 7 right?

    If you're going from 6 drives, in the array, to 7, yes.


    But note if you're running RAID5: Running 7 drives in a RAID 5 array is high risk proposition, especially if the drives have any age on them. You really need to backup data you can't afford to lose.

    Is it okay that it's already added via the GUI?

    In this case, (mdadm RAID) changes can be done in the GUI or on the command line. The outcome will still be viewable in the GUI.

Jetzt mitmachen!

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