Posts by antsh

    I might be not understanding something very simple, but I need a bit of help navigating this "issue".

    I have a 6 port ethernet card in my OMV server. Ports 1 and 2 are configured as a bond, all of that is working fine. The rest of the ports on this card are NOT configured at all in OMV.

    I took a cable and plugged it into port 6 on the above network card, and plugged the other end into my switch (different VLAN if that matters), and the unconfigured interface on port 6 pulled an IP via DHCP (on the correct VLAN). I would expect since the interface is not configured for anything in OMV, it wouldn't actually pull a DHCP lease when the link went up. I actually don't want it to pull a lease in this case (going to pass it through to some VMs).

    I also tried defining the interface in the OMV UI, and setting the IPv4 and IPv6 settings to disabled, but that completely brought the interface down instead of leaving it up but disabling IPv4 and IPv6.

    Is this expected behavior? Is there a way to prevent this?

    Just experienced this when upgrading to the 5.6 kernel. For anyone that comes across this the answer is to install the

    linux-headers-amd64 package. This will automatically install the new headers needed when a new kernel is installed. This way you don't have to manually install the linux headers for your specific kernel every time.

    Got it thanks. This will default these for all shares for my testing. Am I right in assuming that these settings are not currently adjustable at a per share level?

    Hi All,

    Is there an env variable available to change the file create masks and force create modes for samba shares? Right now they are are set as 0664 which works for most things. Also, this will change the create mask and create modes for every share created. Is there a way to change this at a per share level?

    My use case is I am using a share as a mapped drive in windows for a steam library. This mask causes the files that steam creates at installation to be created as non-executable, which of course breaks the launching of games.

    I am looking for a way to create a share that allows files to be created as executable. I am thinking modifying the create mask and mode is the way to solve this, unless there is a better way?


    and in general, is there a way to lookup all the valid env variables that can be set across OMV?

    Thanks, I had no idea you could append configurations with separate files for netplan; this makes it really easy to "enhance" some of the configuration available through the UI.

    Just as an FYI for anyone that sees this later, you need to add the parameters: line to the above, so it should actually looks like this if your interface is bond0:

    transmit-hash-policy: layer3+4

    Thanks as always for the great work. Issue created:

    As a temp fix until this is addressed (as priority dictates, of course) just to add the below to the extra options of every share that isn't a time-machine share?

    fruit:encoding = private
    fruit:locking = none
    fruit:metadata = netatalk
    fruit:resource = file
    vfs objects = catia fruit streams_xattr

    I have a general question regarding the default samba settings in omv5.

    I see that if you have at least one share enabled for time machine then

    fruit:aapl = yes 

    gets set globally and

    vfs objects = catia fruit streams_xattr

    Gets set only on the time machine share.

    The global behavior is as I would expect, but only the time machine share getting the vfs object parameters is counter what I’ve seen from the samba wiki. It recommends to set this on all shares if set on any of them:

    Be careful when mixing shares with and without vfs_fruit. OS X clients negotiate SMB2 AAPL protocol extensions on the first tcon, so mixing shares with and without fruit will globally disable AAPL if the first tcon is without fruit.

    Running testparm actually warns about it as well:

    WARNING: some services use vfs_fruit, others don't. Mounting them in conjunction on OS X clients results in undefined behaviour.

    Is this being done deliberately, or perhaps an oversight? Any harm in manually setting the vfs objects as suggested in the extra options of the non-time machine shares?

    Thanks for taking a look!

    You aren't using one of the containers to access the gui at the time are you? This would happen if you were using the letsencrypt/nginx container to proxy your domain name, and then used that domain name to access omv during the update for example.

    Hi All,

    This doesn't pertain to OMV directly perhaps, as I have a feeling that this is a Debian issue/nuance.

    I am using a MegaRAID card, and there is a monitoring service that is provided as a service called megaclisas-status. When this service tries to start, i get this in the logs and it exits:

    megaclisas-statusd[811]: Starting megaclisas-status monitor: megaclisas-statusddaemon: ignoring unsafe /etc/daemon.conf (/etc is group writable)
    megaclisas-statusd[811]: daemon: fatal: refusing to execute unsafe program: /usr/bin/daemon (/ is group writable)

    Does someone with a bit more linux knowledge than me have an idea how this can be addressed? I am thinking it is a permissions issue, but don't know if messing with daemon is a good idea.

    Can you show your storage mappings for emby for both omv4 and omv5? One thing to note is that omv5 doesn't use /sharedfolders/ by default anymore, so you need to change that mapping. Also, specifically for your emby config files, you have it pointed to the same location in omv4 and omv5 and the same files exist in the same locations for both installs?