• Disks missing after upgrading to 6.7.0


    40foot

    I had to revert to 6.6.7, both my WDC_WD40EFRX were not detected after upgrading, only Toshiba was found, so only 1 out of 3 disks (SSD cache was found too). The file included is after reverting - for analyzing, do I have to upgrade once more, and make the diagnostics under 6.7 again? Can this lead to more severe troubles? Or are there any log-files to be found, where those upgrade errors were written to? Ah, I feel not good about it, please help.

    durosrv-diagnostics-20190513-0050.zip




    User Feedback

    Recommended Comments



    13 hours ago, 40foot said:

    I have exactly the same machine and disabling IOMMU in BIOS/EFI works. What I haven't tried yet (I hate downtimes) is the other option editing the syslinux and reenabling it in BIOS.

    Do you know perhaps what practical implications disabling IOMMU has?

    Or is it perhaps a bad thing to have it enabled since new kernel/unraid doesn't work with it?

    I read a bit on wiki, but... it is out of my scope of knowledge(/interrest).

    Link to comment

    Hi, is this something that Unraid team is working on, or is it something out of scope of unraid team and should be fixed by individual user? Any recommendation or more "official" details as to what is causing this and how to best fix it without risking other problems?

    I am not using any VMs at the moment, but would not like to exclude possibilities if one day I want to, because of some unresolved bug/issue.

     

    I just tried updating from 6.6.7 to last stable, but I am still getting this issue, so I am basically stuck on 6.6.7 

    Link to comment
    10 hours ago, dakipro said:

    Hi, is this something that Unraid team is working on, or is it something out of scope of unraid team and should be fixed by individual user? Any recommendation or more "official" details as to what is causing this and how to best fix it without risking other problems?

    I am not using any VMs at the moment, but would not like to exclude possibilities if one day I want to, because of some unresolved bug/issue.

    This workaround works for some, if it doesn't best bet would be to replace the Marvell controller.

    • Like 1
    Link to comment
    On 6/3/2019 at 9:46 AM, dakipro said:

    Do you know perhaps what practical implications disabling IOMMU has?

    Or is it perhaps a bad thing to have it enabled since new kernel/unraid doesn't work with it?

    I read a bit on wiki, but... it is out of my scope of knowledge(/interrest).

    Disabling IOMMU means that you cannot run VMs with hardware pass-through.   You can still run VMs that do not need hardware pass-through.   Marvel controllers also seem to be prone to dropping disks offline for no discernible reason. 

     

    The issue seems to be a compatibility issue between the Marvel drivers and the latest Linux kernels.  The fix (if one is possible) is outside Limetech's control.   If you want to run VM's with hardware pass-through then the easiest thing to do is to stop using the Marvel controller and switch to using a LSI based one which work fine with current Linux kernels.

    Link to comment

    Thank you for answering @itimpi and @johnnie.black , I will disable IOMMU and try to remember that no hardware pass-trough is possible on the pc. I was planning on having a memory card reader station for cameras, but I will see how and if that would work without pass-through. 

     

    The thing is that I purchased hpe microserver gen10 to use it specifically as a Unraid NAS, since people are using it and it is working great. So the pc is pre-built to be a nas, without options to change pretty much anything on it. I used so much time on choosing right hardware and software for my Nas, and now a relatively new piece of hardware is, as it seams, not supported anymore by linux, and thus by Unraid as well. 

     

    A bit odd that all is working fine on 6.6.7 but not on a new kernel. One could even argue that there IS a solution (running behind me), just that Linus thinks it is not a good one so here comes a "buy a new hardware" solution instead. 

     

    Link to comment

    Sorry for reviving old thread, but this problem is still active for me on 6.9.2.

    It can be fixed by adding amd_iommu=off to the end of the Unraid OS syslinux configuration.
    Both amd_iommu=pt and scsi_mod.use_blk_mq=1 did not work for me.

    My question is: Will I still be able to use IOMMU? Is setting amd_iommu=off the same as turning it off in the bios? Or only for SATA?

    HPE Microserver Gen10 with Opteron X3421

    Marvell 88SE9230
    Unraid 6.9.3

    Link to comment
    44 minutes ago, gezellig said:

    Marvell 88SE9230

    Marvel based controllers are not recommended due to the complete lack of manufacturer support.  Disabling IOMMU seems to help for some users, but it is always recommended to switch to LSI based 

    Link to comment

    For future users with marvell controllers: Make sure you try both the chip specific intel/amd_iommu=pt and also generic iommu=pt because that was what worked for me. I am still able to use IOMMU with other hardware!


    @SquidI understand the marvell controllers are not ideal but it came included with my server and I'd rather not purchase another raid controller and consume my only pcie slot.

     

    Would be possible to automatically enable iommu=pt syslinux flags on future unraid releases if a marvell controller is detected?

    Link to comment

    just in case this helps anyone else. I realised when i was testing this, i had put the iommu=pt but only for the boot without gui. When I first tested this it failed and no disk were shown and it was only because i didn't append it to the end of the gui boot line e.g.

     

    default menu.c32
    menu title Lime Technology, Inc.
    prompt 0
    timeout 50
    label Unraid OS
      menu default
      kernel /bzimage
      append initrd=/bzroot iommu=pt
    label Unraid OS GUI Mode
      kernel /bzimage
      append initrd=/bzroot,/bzroot-gui iommu=pt

     

    So essentially you will need to add it to all the boot options if you want the disks to be shown. Or you can just manually add it if you are not running it headless. 

    Link to comment
    On 6/28/2021 at 6:06 PM, gezellig said:

    For future users with marvell controllers: Make sure you try both the chip specific intel/amd_iommu=pt and also generic iommu=pt because that was what worked for me. I am still able to use IOMMU with other hardware!


    @SquidI understand the marvell controllers are not ideal but it came included with my server and I'd rather not purchase another raid controller and consume my only pcie slot.

     

    Would be possible to automatically enable iommu=pt syslinux flags on future unraid releases if a marvell controller is detected?

    Nice one, I was rebuilding my server and decided to update firmware and after amd_iommu=pt no longer worked on my hp microserveer but after an hour of digging I finally caught this reply and iommu=pt did the magic.

    • Like 2
    Link to comment



    Join the conversation

    You can post now and register later. If you have an account, sign in now to post with your account.
    Note: Your post will require moderator approval before it will be visible.

    Guest
    Add a comment...

    ×   Pasted as rich text.   Restore formatting

      Only 75 emoji are allowed.

    ×   Your link has been automatically embedded.   Display as a link instead

    ×   Your previous content has been restored.   Clear editor

    ×   You cannot paste images directly. Upload or insert images from URL.


  • Status Definitions

     

    Open = Under consideration.

     

    Solved = The issue has been resolved.

     

    Solved version = The issue has been resolved in the indicated release version.

     

    Closed = Feedback or opinion better posted on our forum for discussion. Also for reports we cannot reproduce or need more information. In this case just add a comment and we will review it again.

     

    Retest = Please retest in latest release.


    Priority Definitions

     

    Minor = Something not working correctly.

     

    Urgent = Server crash, data loss, or other showstopper.

     

    Annoyance = Doesn't affect functionality but should be fixed.

     

    Other = Announcement or other non-issue.