NFS crash

1st Official Post

    Hi,



    since several weeks, I observe NFS cashes. My mount is broken, I can't stop nfs-kernel-server, I can't restart it, the only solution is to reboot my server. I have read that NFS was updated recently, but I can't find any procedure to fix my issue.

    OMV7 on Intel N100 DIY NAS.

    I have this. And I use NFS v4 only (Debian 10 and Fedora 40 client).

    OMV7 on Intel N100 DIY NAS.

    Kernel panic


    have alook here

    Linux Kernel panic issue: How to fix hung_task_timeout_secs and blocked for more than 120 seconds problem
    This guide shows how to fix hung_task_timeout_secs and blocked for more than 120 seconds problem in Linux. A panic may occur as a result of a hardware failure…
    www.blackmoreops.com


    might help

    OK, effectively, just before, I have a huge load...


    Before :

    Code
    sysctl -a | grep dirty | grep ratio     
vm.dirty_background_ratio = 1
vm.dirty_ratio = 20

    After :

    Code
    sysctl -a | grep "dirty" | grep ratio     
vm.dirty_background_ratio = 5
vm.dirty_ratio = 10

    OMV7 on Intel N100 DIY NAS.

    did this fix your problem? I've been having this exact issue since I upgraded to omv7. I'll make the changes in the linked url, was just curious. TIA!

    changing the vm dirty settings appears to have fixed my issue so far. Uptime is over 3 days now, where previously it would crash every 24-48 hours.

    this has really only delayed my issues to about 4 days uptime. I've upped resources to 6vcpu (vm running on proxmox) and 14 GB RAM which I think is a bit overkill as performance stats don't show cpu above 60% in the last week or ram usage above 2-3 GB. I've tried using kdump to force a panic, but that won't seem to create a log. the only other change other than omv7 upgrade was moving my *arr stack to omv instead of a seperate VM, but I don't see any mention of docker/container.io in the kernel error below. I did see there were a few updates available for compose/docker, which I've now installed. Anyone have any other ideas?


    including perf stats in case it's helpful



    including the 1st panic trace below.

    I'm getting a similar issue as well. This is a fresh install of OMV7 and not an upgrade from OMV6.


    I noticed that it doesn't happen when I FTP hundreds of gigabytes of data to the NFS share. However, if I'm downloading a large torrent, it will cause it to hang more frequently. It also hung recently without a lot of write traffic, so I'm a bit confused. I've tried setting the dirty attributes and it does delay it, but it still happens.


    Edit: Additionally after I initiate a reboot from the terminal, it'll take awhile and fail with this


    I also crash, OMV and Qbittorrent on a separate VM, in Proxmox. OMV shares a drive ( HDD passthrough) via NFS.


    This occurs with torrents over 10-15GB.


    I would like to add that the problem has been occurring for about 3-4 months, earlier it did not occur and I did not change anything in OMV except updates.


    It is annoying, on top of that I can't conterner with OMV to restart. Only have to force terminate it.




    wPtwpNm.png


    J7G4T6B.png

    Quote from Camis

    I also crash, OMV and Qbittorrent on a separate VM, in Proxmox. OMV shares a drive ( HDD passthrough) via NFS.


    This occurs with torrents over 10-15GB.


    I would like to add that the problem has been occurring for about 3-4 months, earlier it did not occur and I did not change anything in OMV except updates.

    Do you happen to know which update was where you started seeing the issues?

    Quote from tytek

    Do you happen to know which update was where you started seeing the issues?

    Unfortunately I don't know, for testing I put up a VM with an earlier version 6 of OMV.


    And it runs ok does not crash.


    6.5.0-3 - ok

    Latest 6.9.16-1 - crash

    Latest 7.4.4 - crash

    Hi, I've experienced the same issue on last month. Yesterday, after finding this thread, I've changed the two values:

    Code
    vm.dirty_background_ratio = 5
vm.dirty_ratio = 10

    I've a host (RPI) where there's an instance of qBittorrent working over NFS on my OMV (it's running on a minipc with 16GB of RAM). I don't know from which system's upgrade it has started happening this issue.


    If I can share some more informations please tell me, thanks!

    • New

    Same problem here after manual upgrade from OVM6 to OVM7.


    After a large workload, NFS hangs on the client and server.

    The server cannot be rebooted as some taks are still running (as mentioned by tytek).

    I run my OMV in a VM on Proxmox with 2 cores and 2 GB RAM on a ZFS filesystem.


    Setting the value of vm.dirty_background_ratio = 5 and vm.dirty_ratio = 10 didn't help me.


    Here is a short excerpt (sudo journalctl -xe):

Participate now!

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

Register Yourself Login