Beiträge von Gent001st

    Hey again,

    you are right, procfs is populated with ipv6. I made few more tests and problem is that I added ipv6.disable=1 after the "--" (double dash) in the kerne params which was treated as filenames. If I remove double dash, everything is working fine so that was completely my fault.

    Once again, thank you for your help ;)

    ryecoaaron - make sense with /etc/issue


    I did once more fresh install and I can see before reboot:

    1. ipv6.disabled=1 in /target/etc/default/grub and /target/boot/grub.cfg
    2. /target/etc/network/interfaces is correct
    3. /target/etc/nginx/sites-available/openmediavault-webgui contains IPv6 listen directive

    So it seems config files are just not regenerated on first boot and are kept from installation so I should do it manually with omv-salt stage run deploy when changing boot options - my fault, I expected to be generated on first boot.

    ryecoaaron - thank you for the wiki link and info, I will check that ;)


    Krisbee - btw: did you try install with Openmediavault_7.0-32-amd64.iso? I installed OMV6 few times last week without any issue, so that is why I thought it might be a bug in v7 (all tested in proxmox VM). Only I can say is that "strange" netconfig (with ipv6 and virbr0 mentioned above) comes from file /target/etc/issue after finishing the installation and before reboot. That file become /etc/issue after the reboot and gets populated with correct network settings after few seconds.

    Hi,

    I realized issue with IPv6 after OMV 7 fresh install (Openmediavault_7.0-32-amd64.iso).

    I booted up the installation ISO with parameters "biosdevname=0 net.ifnames=0 ipv6.disable=1" and successfully installed the OMV7. Right after the installation, the webGUI was unreachable and after a bit of investigation I found that nginx does not start.


    Problem is that nginx configuration file is "generated" with IPv6 (which is completely disabled on boot by ipv6.disable=1.):


    /etc/nginx/sites-available/openmediavault-webgui

    Zitat

    listen [::]:80 default_server;


    It can be easily fixed by deleting that line from nginx's configuration file but I think there is some race condition in first-run scripts and therefor IPv6 detection does not work correctly.


    Right after the first boot I can see message (see attachment) on the console for ~5seconds from (/target)/etc/issue with "unconfigured interfaces": (message and IP addresses are always the same, I tried fresh install multiple times):

    Then (after ~5seconds) console get refreshed and correct interfaces configuration appears, so it makes me thing first-run script (including IPv6 detection) run before the system is completely configured and therefor all daemons gets (mis)configured with IPv6:


    journalctl | grep -i "Address family not supported"


    I can fix it myself, just want to let you know, what I have discovered.


    Cheers and thank you for your hard work ;)