Posts by Swampy

    OK, I screwed up by not deleting incompatible plugins before upgrading from OMV 3.x to OMV 4.x. I just assumed the installation script would either disable incompatible plugins or remove them with a warning message. Silly me. It did neither, which resulted in an annoying "Invalid RPC" error message almost everywhere in OMV. So since my server was barely configured with OMV 3.x, I decided just to do a clean install of OMV 4.x. What could possibly go wrong?


    The installation went fine until it came time to configure a boot drive. For OMV 3.x, I was using a 16GB SLC USB drive with the Flashmemory plugin. It seemed to work fine. Then, when installing OMV 4.1, right after entering the root password, I got error messages: "Failed to partition the selected disk" followed by "No root file system is defined." At first, I also saw the installer ynoticed earlier OMV files from the 3.X installation on the drive.


    So the first thing I did was pop the USB drive into my iMac, fire up Disk Utility, and erased the drive. After trying this a couple of time without it working, I set the erase level to 7, which erases existing data, writes over it, etc. Still had the problem.


    Then I read this page, this page, and this page. I tried all three suggested solutions: ryecoaaron's 7/10/14 suggestion to zero out what was on the drive, Raj Kumar's 5/12/13 suggestion to use GParted to create new swap and system partitions on the drive, and ryecoaaron's 12/10/16 post, which was like his earlier one but increases the number of zeros by a factor of 10, from 1000 to 10000. I tried each of them, but still no luck.


    One thing I don't understand, is that the error messages mention fixing the partitions from the partition menu. If I escape back to the menu of everything one can do during installation and select the partition item in the menu, the system throws another error, saying there's no root directory. But if I'm using a different USB with the OMV installation iso, there's obviously a root directory there. And isn't the purpose of a partition utility, at least sometimes, to create partitions? So if you're preparing a blank drive, of course there won't be a root directory before you've configured the drive. What am I missing?


    Could someone please help me with this? ;(


    Thanks.

    I made the mistake of not uninstalling 4 plugins (shellinabox, usbbackup, route, and netatalk) that are not crucial to me but I had installed when I set up OMV 3. I tried uninstalling them after the upgrade but still get the Invalid RPC message.


    Is there any way of getting them off the system now, so the error message stops coming up?

    Clonezilla is for making disk images of unmounted drives on the local machine.


    Urbackup is for automatic full and incremental file backup (Linux or Windows) or image backup (Windows only) on remote live machines.


    Those are totally different purposes.

    Excellent, concise explanation. Thank you.


    From what you say, is it correct to infer that urbackup is also not for use with Macs because:

    • You say just "Linux or Windows"?
    • It duplicates what Apple's Time Machine does?

    Hi folks,


    Very new to OMV and a bit confused about how to use the Symlinks plugin. I've created a 2-drive pool, Data1and2, that according to the UnionFileSystems plugin resides at /dev/sdb1. I want to create a new directory, Docker_Images, at the pool's root. Then I want to create a symlink to this directory, and then use the link as a path in the Docker plugin.


    Here's what I tried:

    • Symlink plugin > +Add.
    • In the Source box, click on the folder icon and navigate to /dev.
    • Look for sdb1 or even sdb*. Nothing beginning with sdb shows.

    What's wrong? ;(


    Also, although I haven't gotten that far, once I locate the root directory of Data1and2, can I create a folder there by using the Symlink plugin and simply adding the folder to end of the Source path? Or do I have to create the folder some other way? What about ownership, group membership, and permissions for the newly created folder? Is the only way by using the CLI, or does OMV have some tool that can do this?


    And once the link is created in the Symlink plugin, is using the link a simple matter of copying the Destination path and pasting it in another plugin's text box? Or must one use some other way?


    My goal is to use a symlink in the Docker plugin so that if I decide to move the images somewhere else, all I'll need to do is redefine the link.


    Thanks for your help!

    Hi folks,


    I have some general questions and home some of you will take pity and answer them. I've been working for months, off and on, to get a NAS up and going. But I'm new to Debian, OMV, and Docker, so there's lots to absorb. So far I've installed and set up OMV 3.0.98. I'm not really using the NAS for anything yet, and all I really plan to use it for is: storing media, media service (Plex Pass version of Plex, Kodi, and PlexPy), WebDAV server, and TimeMachine backup. I have read the Docker documentation, and this entire thread (all 3 years, 43 pages, and 848 previous posts). Still I'm feeling quite clueless.


    As a first step, I've decided to work on getting Plex going. But I have some basic questions:

    • Before going to the trouble of finishing setting up the server, should I install OMV 4.x instead of 3.x? The OMV Downloads page has OMV 4.0.14 available, and it does not have the usual suffixes (.beta, .rc, etc.) for versions still being tested. Several of you say you are using OMV 4, and the OMV4/Docker/Docker GUI seem better integrated. OTOH, I've not seen any announcements that OMV 4 is officially released. If I set things up under OMV 3, would I have to just do it over again with OMV 4? OTOH, should I trust OMV 4 now?
    • Lots of posts in this thread show windows labelled "Source Code." But much of what's shown looks like what I'd ordinarily call console output. Is this just a matter of nomenclature? Or there there more here than meets the eye?
    • Whatever these windows are showing, how does one create them? Does OMV have a tool to capture the image? Or what?
    • Logging in via SSH and exploring the file system, I get the impression that the system expects home directories to be located in /home. But looking around the OMV GUI, I come across Users. I've seen both on *nix systems. Could someone straighten this out, please?

    I realize these questions are mainly about OMV and its relation to the underlying OS. But they've come up as I've been plodding my way trying to get Docker going with Plex. Without knowing the answers to them, I don't think I can make intelligent decisions about Docker going forward. So I hope you'll help with this.


    Thanks.

    Images and containers use most of the space. Maybe you have unmapped data volumes there, those cannot be managed from the plugin.
    You don't need to use the os drive btw, the plugin can pass the argument to the daemon to change the default docker folder (use a shared folder) or you can use a symlink to a data drive(my preferred choice).

    The symlink approach seems best. Can you give a bit more explanation of how to do it?


    I'm very new to OMV and Docker, so all I can do is guess. I would guess using the CLI to navigate somewhere. But from /, I'm not sure where to go. I have a shared folder, "Data 1 and 2," and I'd like to create a directory (folder) called something like "docker_images" there. How does one navigate from / to the base of a shared folder? Once there, I'd mkdir a new directory and chown & chgrp it to "dockeradmin" (the docker owner) and "users", and then I'd create a symlink. But this is all in a SSH window. Do you then just cut and paste the symlink's full path into the Docker GUI?


    Also, if down the road you wanted to change the location where Docker images are stored, how would that work? Keep the symlink but make it point to the new location? Reconfigure the default folder in the GUI to point to the new place with a new symlink? Something else? And would you have to move all the images from the old place to the new one, or would Docker automatically download new images when it sees they're missing from the (new) default location?

    A few minutes ago I submitted a question to the forum and got an error message. The error code was: bb1f8e5c159d0658be57ac43a24ab00acb3bcd5c.


    The message said to forward the code to the site administrator, by which I assume it means the administrator of this web site. I can't find an email address or form to contact the site's administrator. How does one send an error code like this? To where?

    I'm setting up an OMV NAS to replace my Netgear ReadyNAS. Currently, I've been trying to configure SSL, something I've never done before.


    From what I've read, an SSL certificate is associated with a domain name. This presents my first problem. Currently I use two free dynamic DNS host names from noip with the ReadyNAS. But both are used solely for WebDAV and point to a share devoted to WebDAV access. I expect they'll be used the same way on the new NAS, and I don't really need a domain name for points higher up in the file system hierarchy. That is, unless I do need one to implement SSL for OMV. Right now, while I'm still configuring the new server, using its local network IP address is perfectly fine. To configure SSL do I need to associate a domain with the NAS drive's IP address, or can I just use the IP address without a domain name and still use SSL?


    Just to be clear, noip provides several domain names to choose from, but the user must provide another name to construct a host name. This really the only thing the user "owns." For example, noip might offer a domain name of "foo.net," and the user could then use "bar" to create a host name of bar.foo.net. If the user were to use this as the domain for the SSL, it would not strictly speaking be a domain name. OTOH, if the user only used "foo.net," then this would be a domain shared by hundreds of other, anonymous users.


    FWIW, I do realize that once the NAS is up and running, if I want to administer the NAS drive from outside the local network, I'll need a dynamic DNS because my ISP uses DHCP to assign IP addresses to my router.


    Second, when I've looked online for a free SSL, I've seen sites that also require an email address associated with the domain. But I don't run an email server on my NAS, so even if I got a domain name for it, I still wouldn't have an email address using the domain. What's the story with this?


    Third, thus far I've had no luck using SSL certificates in OMV. I've tried:

    • The self-sign certificate method described here
    • and the LetsEncrypt plugin method described here. [It asks for both the domains to use and the root directory of the Internet-facing webserver. I have no idea what directory I should use. And for now, I'd prefer the IP address over a domain.]

    Maybe these haven't worked because the NAS drive doesn't have a domain name. But even if I were to plan to give it a domain name, I wouldn't want to do so yet because the machine is not properly configured and still has several security issues. I want to configure it before making reachable from outside my local network. What's the best way to do this?


    Thanks.

    If two different NAS devices on a single local network are both being used as WebDAV servers, how does one set up the ports?

    Hi,


    I'm new to OMV & Docker, and have three questions. One relates to limitations using docker with OMV. The other two relate to docker conventions. I'm hoping some of you out there can help.


    1. I'm confused about the Docker base path (DBP). My OS is stored on a relatively small (16GB) USB drive, so I want to keep things like Docker files on a hard drive instead. I tried using the GUI to set the DBP in a shared folder, but the GUI would not let me enter subdirectories in the path.


    Then I read Subzero79's Sept. 3 comment here, which seems to imply that one should not use the plugin's GUI to reserve a shared OMV folder for the DBP. He suggests instead using CLI to create a symlink from /var/lib/docker to the folder. If one does it this way, is there any problem with putting the DBP in a shared folder spanning multiple devices? Also, /var is used for many things. So if one wants to keep those many things off the system drive, wouldn't the symlink connect just /var to the shared folder, with lib/docker beneath it? In general, could someone please explain what the problem is with using the GUI and multi-device shared folders? And what's considered best practice for locating the DBP off the system drive?


    2. As a general rule, I try to follow conventions whenever possible. For example, on *nix systems I always use ~/bin to store my own programs and shell scripts because this parallels the systemwide conventions for /bin and /usr/bin. So if /var/lib were to stay on the system drive, where would the docker directory go if it's off the system drive? One place, suggested in #1, is to create a symlink in /var/lib to docker/ on a different device. But what if someone didn't want to mess with symlinks and the CLI? Suppose they just wanted to use the GUI to set the path; what would they call the docker directory? Would it be docker/ at the top level in the shared folder? Would it be var/lib/docker/? Etc. What if different users were going to maintain their own docker images and containers. Would ~/user/var/lib/docker be the DBP for each of them? Or is the nature of docker such that there should only be system-wide storage of containers and images?


    3. Similarly, this guide recommends creating a special user to own docker folders, etc. The guide leaves the name up to the reader but uses "dockeruser" to stand in. Is this the typical recommended practice? If so, is there a conventional name for this user, the same way "root" is by convention the name of the super user? I would also think the plugin would automatically create this user when the plugin is installed. If creating a special user to "own" docker is not the typical recommended practice, then who should own the docker files? Can a system administrator be the owner?


    Thanks.

    Thanks, drinks.


    The problem I had is that my BIOS only allows choosing boot order, with USB being one of several choices for the first item in the boot sequence. But I had two USB devices connected to the system, and the BIOS isn't smart enough to figure out which one to use as the boot device. (This is on a Supermicro X11SSH-F MB). I had assumed it would be smart enough to use the bootable USB, but it wasn't. IIRC, the motherboard has 10 USB ports, and AFAIK there's no way to specify individual USB ports in the boot order.


    Also, what you say about 3.0 vs 2.0 makes sense. Very likely the OS kernel plus a few frequently used services are loaded into RAM at boot time. So once the system boots, there's not much i/o on the boot drive that would affect performance.


    But I'm using a MX-Technology 16GB MX-ES 3.0 USB drive because its memory is SLC NAND and therefore more reliable than the cheaper MLC technology. For the slight cost premium, this seems a worthwhile tradeoff for a system drive. Besides, the MB has one internal USB port, intended for hosting a boot device, and it happens to be 3.0. Since the device is 3.0 anyway, there's really no other choice but 3.0. But any speed improvement from using 3.0 is just coincidental: SLC NAND is the reason for using the device, which just happens to be 3.0.


    Nonetheless, I've got the flash memory plugin running too. Over-engineering is the way to go.

    Just an update.

    • Rebooting got rid of the PMS item in the sidebar, but the XPath error message still sometimes crops up in unexpected places.
    • {Uninstall plugin > reboot > reinstall plugin > reboot} does not fix the problem.

      • The first time I tried to install the plugin the installation seemed successful, but the PMS did not start. Following suggestions in the forums, I then tried uninstalling and reinstalling the plugin. But, except for the initial install, it does not install without errors.
    • I think my next step will be to try installing a linux version of PMS in docker.
    • But before doing so, I'd like to clean up all traces of the unsuccessful Plex plugin installation. Do you have any suggestions for the best way to do this?


    Thanks!


    I think for now I'll try using the WebDAV plugin. The last release of SabreDAV is recent, so my concern is more long-term. Maybe someone else will pick up supporting SabreDAV or the author(s) of the WebDAV plugin will find a substitute before the current SabreDAV becomes too obsolete.


    In the meantime, I'll learn about Docker and use it to get Plex Pass version of Plex up and running. I tried the OMV Plex plugin but had lots of trouble. And since I want to use the Plex Pass version anyway, the plugin was just a stopgap measure.


    With this Docker experience under my belt, I'll revisit WebDAV.

    Apache is a no go in the host at least in the form of a plugin, Omv used to have apache as http engine before, but the performance was much better with nginx.
    My opinion would be to move this yo docker. You'll loose pam login.

    Thanks for the advice & insight, subzero.


    I understand docker containerizes apps, but I've never used it. I see there's some discussion of it in this forum, and even a GUI plugin. So I'll put off questions about docker until I've had a chance to climb the learning curve.


    But you suggest moving "this" to docker. Exactly what "this" do you mean? Apache, some other WebDAV-providing FOSS http server that's not dependent on SabreDAV, or something else?

    OMV's WebDAV plugin relies on SabreDAV. In April SabreDAV's developer stopped supporting SabreDAV. Since no other developer has stepped forward to take over support, this implies SabreDAV is essentially petrified in its current version. In turn, this implies OMV's WebDAV is too.


    WebDAV is built on http, and OMV uses NGNIX for http services. According to Wikipedia, NGNIX has "a very limited optional WebDAV module" (which is not automatically installed). So even if OMV started using NGNIX WebDAV, it's not clear that this would meet users' needs.


    OTOH, according to the same Wikipedia article, Apache provides WebDAV modules.


    So this is appears to be a case of one domino affecting others. Moving forward, what's the best way for OMV to provide WebDAV services that will continue to be supported?

    Hi,


    I'm just starting to configure WebDAV on my system and have three questions:

    • I notice Services > WebDAV requires choosing a shared folder but only allows one folder to be chosen as the WebDAV folder. What's the recommended approach to set up OMV if one wants to allow several WebDAV folders to be used for different purposes and accessed independently?
    • My local network already has a WebDAV server. So to bring up the OMV server as a second WebDAV server, I suspect I'll need to use port forwarding. Is this so, and if so, what ports should I use?

      • The ports currently in use for the other server are: 443 -> 443 and 80 -> 80.
    • It would be helpful to be able to learn about these kinds of things without having to ask every little question on the forum. So somewhere is there a guide to setting up WebDAV on OMV or some other documentation specific to the WebDAV plugin?


    Oh wait. I just started to work with the OMV WebDAV plugin and discovered that it uses SabreDAV as the underlying technology. So never mind the earlier question.


    But since it rides on SabreDAV, does the OMV plugin just do WebDAV or can it supply other *DAV services too?