Beiträge von technicali

Zitat von ryecoaaron

It really depends on what you are trying to do. If you have an ip set in network-manager and then you set the same ip in the OMV web interface, obviously it will keep doing the right thing but both will be doing it. If you have it set to dhcp, it is possible you could get two dhcp address.

OMV uses netplan and systemd-networkd.

Neither knows nor cares about the other one. It is possible to control network-manager with netplan but OMV does not.

Somewhat. They are both affecting the same network adapters but that is where it ends.

Yes.

You really don't want to continue to use both. Pick one.

Alles anzeigen

Thanks for answering each of my questions. It makes a lot more sense now on how with network-manager installed, there would effectively be two systems controlling the same network adapters. The whole thing is fascinating to me as I get more familiar with Linux, learning the ropes and try to setup the "ultimate system", currently with OMV running the show so appreciate you clarifying things.

I looked up Netplan and was interested to see that it supports both Systemd-networkd and NetworkManager as "renderers". My brain can't help but wonder if the Netplan files OMV generates/uses (abstracted network settings in YAML format from the look of it?) can be "applied" to network-manager. I suppose I've answered my own question - they can't since then it could/would overwrite settings that Home Assistant needs. Just thinking out loud and something that's clearly a pointless pursuit in practice. I certainly don't want to take up any more of anyone's time with this as the answers so far are clear. Sometimes I need to know when to stop! As mentioned, I'll now almost certainly keep Home Assistant on my Pi4.

Thanks again!

Zitat von geaves

When I read this thread initially and after ryecoaaron post I decided to do some research, Network Manager is only necessary when using Supervisor as this uses Network Manager on the host to configure networking for add-ons, information is here

Surely if Home Assistant was being used in Docker, then to use Supervisor one could deploy Macvlan, by using a Macvlan config within the Docker container configuring networking for Supervisor and add-ons would be straight forward.

Interesting. I looked up Macvlan as I've never seen it mentioned before. Certainly looks like a possible route to take in setting things up although the learning curve to use it is prohibitive for me at the moment. I think overall, I'll probably keep Home Assistant on the Pi4 it's currently running on since it separates out that fairly important functionality to a dedicated device.

What I'm still curious about though is what actions I can take on OMV with network settings changes using the OMV UI before breaking something with network-manager installed. I've changed the system IP address with no obvious issues.

I'm not sure exactly what OMV uses to manage and set networking settings and how this responds to network-manager being installed? Do they access and change the same centralised system settings? Are they both active at the same time? Are the potential problems mostly from settings being changed by network-manager which OMV isn't aware of and vice versa? In which case, I'm guessing a misconfiguration could be fairly easily fixed. I'm not trying to make a bad idea work here - I'm just curious as to the technicalities in how OMV and network-manager could conflict.

I've got OMV installed in a VM working very well so I'll do some experiments with network settings (with snapshots!) and see if I can break things and figure out why.

Zitat von BebeMischa

As i read this, and more topics, it has been a good choice to keep OMV living on one machine and HA on another machine. Next to each other, as a good friends with each it's own life

Very probably what I'll do! But it is appealing to have everything running on one system.

Zitat von ryecoaaron

I am very curious why something in a container cares about network-manager outside of a container...

Other than the fact network-manager is terrible on anything other than a laptop, you can install network-manager but don't use the network tab in OMV. If you use both, it will configure things with netplan along side of your network-manager.

As for the using packages from sid, you will learn soon why you shouldn't do that. It is called unstable for a reason.

Thanks for explaining and letting me know about the potential issues with network-manager. I think Home assistant uses network-manager installed on the host system to manage the networking settings it uses and requires even though it runs as a collection of Docker containers due to it's functionality on the network.

If I do run Home Assistant on Docker on OMV (not sure I'll go this route yet), I'll have to think carefully about how to manage network settings. As I said though, I updated the IP address of OMV from within the OMV settings after I had installed network-manager and it seemed to handle it fine. I wonder if there's a way to set this up so that Home Assistant is happy with network-manager installed but it doesn't interfere with OMV networking settings - probably not!

I'm not sure I'll use Home Assistant running on Docker on OMV - it might be better to keep it running on my Raspberry Pi4 as a self contained dedicated system but I did like the idea of having OMV basically run everything I needed all in one place.

As far as I'm aware, I'm not using packages from sid so no problems there.

I've seen this thread here: Network-manager where someone asked about installing network-manager in OMV and the reply was very clear: "Installing network-manager is a bad idea since OMV manages the network settings. Installing packages from sid is generally a bad idea as well."

At the moment I have set up an OMV installation in Virtualbox and set it up with Portainer and Docker, OMV Extras and a couple of attached virtual hard drives for testing and it's all working very well.

As an experiment, I wanted to see how a Home Assistant docker container would run on OMV so I set it up using the manual "supervised" installation commands.

Part of the installation is that it requires network-manager so I took a snapshot on the virtual machine as a precaution and installed network-manager using sudo apt, bearing the above warnings in mind. To my surprise, everything seems to work fine after install. I even changed the IP address of the machine using the OMV interface and there are no apparent problems.

Now I have no doubt that installing network-manager on OMV is generally a bad idea since I'm guessing it uses it's own networking system/settings but wanted to ask if anyone could explain more abot what kind of issues it would cause and specifically and if there are ways to meet the requirements of Home Assistant with network-manager installed while also keeping OMV problem free? I don't notice any issues with my test installation now, but I suspect problems could arise in future with network-manager installed?

Zitat von donh

You can install proxmox on a minimal debian install. That way you can set the disks up the way you like. I think setting up omv in an lxd container would work but I wouldn't do it that way. If you try and it works fine let us know.

I use zfs mirrors for important data. Remember mirrors are not backups.

Thanks for the suggestions and info. I'll hopefully have time over the weekend to give it a go. If I manage it I'll update here

Zitat von donh

I run omv as a vm in proxmox. I don't use docker containers and only a few plugins. The omv uses very little system resources. It makes it easy to snapshot omv before an update and roll back if there is a problem. OMV makes sharing disks easy and proxmox makes it easy to increase the size of the virtual disks if needed or move them to a new disk.

Try it you will like it.

as for docker again I don't use that but proxmox does do lxd containers natively. They work fine.

That's really interesting. I did get Proxmox installed but couldn't figure out how to set up storage properly from within the interface as I wasn't sure how LVM should be configured or if I should use ZFS and mirror my two 1TB HDD's. Also realised that Proxmox used my whole SSD for system but I'd have liked to have used my SSD to store my Docker containers. I suppose if Proxmox can be run from a USB drive that might solve that problem easily.Main issue I had is that since Proxmox is fairly obscure I found it difficult to find any direct answers/guides as to how to set up hard drives/storage within the interface for either VM's or containers (is there an OMV LXD/C container?). I've read the documentation but am looking for practical guides/tutorials. Is the LVM the way to go? (I'm not 100% sure what the LVM is or how it should ideally be used.

Would be interested to know how you've got your storage set up in Proxmox and how is it configured in relation to OMV?

Zitat von crashtest

RE thumbdrives:
It's just dirt simple to clone them and with one working clone, one backup, and one archive copy (the way I do it) it would be difficult to mess that up. You'd have a back out option. (I do label them, to prevent confusion.)
For speed of cloning I'd go for the 32GB size. 64GB would last longer (wear leveling, also mentioned in the guide) but the cloning operation would take up to twice as long. If I used 64GB drives, I'd test Win32DM's option, "Read Only Allocated Partitions" to see how that works. (I have yet to try that option, but if the source and destination pass verification, that's what matters.)
BTW: The image file itself is backup. With a date in the file name, if it's saved off to the side, you'd have a history of images to recover to. On occasion, I'll run an MD5 on a file, then compress it with 7zip and save it as an archive. The file is another recovery option.

On using SSD's to boot:
I don't understand the purpose behind partitioning a boot SSD. If you're going to use an SSD to boot, if you want ease of recovery, you'd need two. Otherwise, if you partition an SDD into two partitions, if the device itself fails, you're dead in the water. (No backup.)

Alles anzeigen

I've just reinstalled OMV on my SSD and am looking into getting a clonezilla backup of the system partition. Your idea about the USB pendrives sounds great and I'm going to order some to try that approach too as it does simplify things a lot. Good point about the SSD and if it dies the "backup" goes with it. I think I need more sleep instead of tinkering for hours at a time with this stuff! Thanks for your suggestions they've been very helpful.

Zitat von crashtest

There are many ways to backup OMV (as in the OS) but the easiest I've found and the absolute fastest to restore is, using and cloning thumbdrives. ( Here, on page 75 in the current version.) While it's an offline process, it's simple and effective and would allow you to gracefully back out of a configuration mistake. (As you're well aware, Dockers can be finicky.)Lastly the Docker plugin allows for relocating Docker images under settings. Moving them from the boot-thumbdrive, to a data drive, is a good idea. Due to the potential for constant activity, USB thumbdrives is not a good place for most Dockers to live. Also, if you go with the thumbdrive option, you'll need to install the flashmemory plugin. That's in the guide as well.
______________________________________________

If you want to use SSD's to boot, there's another option that would work along similar lines but you'd need 2 SSD's installed. In OMV-Extra's there's an option for installing Clonzilla. OMV allows a one time boot to Clonzilla. Clonzilla allows an SSH connection, useful for a headless server, where you could clone your boot SSD to the 2nd SSD for backup. (The choice of booting to one or the other could be made in BIOS.) This is a choice worth consideration.
______________________________________________

RE Proxmox:
I have Proxmox setup, but don't have an OMV VM running on it. (So I can't provide a performance comparison.) This is a pretty elaborate step to take, for the sake of being able to take snapshots of an operating system. Proxmox adds layers of complexity in configuring a VM guest (OMV) and in configuring the storage back end. If the only purpose for considering Proxmox is backing up OMV's OS, I'd go with either of the two above (Thumbdrives or Clonzilla with SSD's).

Other forum users may have other opinions or ideas.

Alles anzeigen

They're some really interesting ideas - thanks! Using USB to boot from seems like it's probably the way to go and I'm just trying to figure out whether to go for a 32GB pair of drives or 64GB since they're fairly cheap. I've read the PDF guide linked which is very useful and see that it mentions Win32diskimager which I'm familiar with. I do still think that it would be a hugely beneficial addition to have a GUI for backing up to images to a storage device from within OMV but we do have to work with what we have, maybe this will be considered for future releases?

For the SSD idea - i'm wondering if it would work with one SSD partitioned down the middle? My SSD is 120GB so it would be two 60GB partitions. I guess in that case I'd need to be able to edit the GRUB menu? I'm not sure exactly how to do that or even if a partitioned SSD would work?

I tried installing and setting up Promox and realised what you said in that it's overkill for what I'm trying to do - looked interesting though even if I stumbled at the hurdle of trying to figure out how to set up storage.

So I've just bought new hardware for a server intended to run Openmediavault as a media server running Docker containers. It's almost the same hardware as the build here with an Asrock J4105 as the base but with 8GB ram, a 128GB SSD and two 1TB 3.5" hard drives.

The main OMV installation was going to be installed on the SSD which I'd then repartition with Gparted GUI to give some of the SSD space as fast config "storage" for configuration files mostly related to Docker containers, particularly Homeassistant. One of the 1TB HDD's would be used for media file storage and the other would receive backup/copies of files from (I hadn't fully figured this out yet) the first HDD using rsync or similar (rather than raid). I'd then also use the Duplicati plugin to upload files to Backblaze.

I've been tinkering with OMV with Docker containers (which are fantastic) on an old Intel Atom D525 build for a while and found that the Docker side of things in particular can throw up unexpected outcomes, errors, "ghost" containers appearing etc which combined with the fact that I'm inexperienced and getting used to how it all works and the linux commandline even though I do have a little beginner experience, meant that I really needed a good backup strategy which I hadn't figured out. Despite this, I'd got together a decent system configuration after hours and hours of quirks probably partly related to hardware issues.... and then my hardware died...

A new system was needed and I eventually got hold of a J4015 motherboard (Intel is having supply issues and these things are like Hen's teeth). However, after searching the forums here, I just couldn't figure out the essential task of a reliable way of fairly easily backing up the entire system for when I wanted to test different settings or modify Docker containers or something went wrong etc. I'm used to using Macrium reflect in Windows to very easily create a differential image of the system drive which can just as easily be restored in case of any issues.

Then I thought of virtualisation and initially ESXI and the idea of running OMV virtualised. This would give me the ability to snapshot and restore the system drive easily but I'm not so sure how setting up the drives would work although I could figure that out. From there I remembered Promox which would kind of make more sense since it's Debian (as with OMV) Linux based. That added questions however since apparently you can run Docker directly from the Promox GUI (I've seen Portainer which I think does/helps with this). At that point, I'm starting to wonder if it makes sense to use OMV at all. If i can just use Portainer to run and manage Docker containers then I could conceivably also just set up a SAMBA server within/from Promox which could serve the files accumulated by the Docker containers? Also there was the question of performance which was important. The J4015 is fairly powerful for what it is with Hardware Virtualisation support, but it's not a fully fledged computer in terms of speed and I couldn't find a definitive answer about the performance hit of running OMV virtualised.

So I'm wondering if anyone has any thoughts on:

• Is there a fairly straightforward method of backing up/snapshotting an OMV installation which is installed directly to hardware, preferably incrementally, with a GUI and which can similarly be restored? I've seen many posts about this question but all of them are pretty opaque in exactly how restores are accomplished and usually refer to Clonezilla or command line tools like DD. The DD based backup tool provided in the OMV GUI (maybe OMV-extras) interface isn't very configurable from what I can see (how to backup only used parts of the disk and enable compression) and again - how would one easily and reliably restore from a backup this way? I think the OMV FAQ is crying out for a super clear and step by step way of creating and restoring OMV installation backups as many people ask the same questions but get answers assuming a "figure it out yourself, here's the linux commands you need to use" approach.
• If backing up and restoring an OMV installation (with snapshots or incrementally ideally) simply isn't an easy thing to do and necessarily requires either the use of existing tools "as-is" or command line expertise - fair enough. In that case - has anyone got experience running OMV as a Promox VM and ALSO on bare metal and knows the performance difference/penalty of doing it that way? How should storage be set up in this case? Does OMV need to be configured/updated with drivers etc for optimal performance on Promox? Are there any other things to consider?
• If i'm going this direction - then would it even make sense to run OMV inside a Promox VM? Could I just as well use it to run Docker containers and then also set it up to serve files over the network, thereby sidestepping the overhead of running OMV virtualised? - although visualising OMV would bring the previously mentioned benefits of being able to snapshot the WHOLE VM, including Docker setup and configs.

There's quite a few ways of doing this and I'm guessing drawbacks/benefits to each approach - any insight/feedback/advice would be great. Thanks

Zitat von gderf

Deluge supports two types of SOCKS5, with or without authentication.

My Streisand server has detailed instructions for setting up a Socks connection but doesn't specify anywhere what the "username" required by Deluge needs to be, if there even is one. I've tried without a username entered and also without authentication. I've tried to find out what Streisand sets as the default username for Shadowsocks and/or how to get Deluge to ignore the username but haven't figured out the correct settings. If i figure it out I'll come back and update but I'm thinking now to reinstall my Streisand server with Tinyproxy support and try to use that. Shame I couldn't get Shadowsocks to work though...

UPDATE: I couldn't get Shadowsocks to work which is a shame but I finally got the haugene/transmission-openvpn Docker container to work with my Streisand server. Was easier than expected in the end I just needed to copy the ovpn file to OMV and specify it correctly in the Environment variables which I'd consistently messed up previously. I did manage to run the configuration script to edit my ovpn file as specified in the documentation but turns out it didn't edit anything at all. If anyone wanting a similar setup needs instructions on how to get it working let me know and I'd be happy to help.

Zitat von gderf

Consider using Deluge for torrents with a SOCKS5 proxy instead of a VPN. The setup is painless and it just works.

Turns out I spoke too soon and there's a bit of a snag. Deluge is asking for a username for the Shadowsocks with Auth option and my Streisand server doesn't seem to specify one as far as I can see. I'll check again tomorrow but fairly sure. No amount of googling has answered the question of what to put in the username field Ive seen that tinyproxy might be an option but I'll have to check.that tomorrow too to see if it's a standard username/password combo. Great idea to use Shadowsocks but as has been the way of things today, I've been scuppered at the final hurdle.

As an aside, I've seen mention that Shadowsocks doesn't handle UDP which could be an issue with using it with Deluge?

Sent from my Mi A1 using Tapatalk

Zitat von gderf

Consider using Deluge for torrents with a SOCKS5 proxy instead of a VPN. The setup is painless and it just works.

Wow - all those hours I spent trying to figure out the workings of Docker and VPN configurations and it was potentially so simple all along (of course!) I'd still like to learn how to get the VPN thing going eventually and learning Docker is really interesting but this is a fantastic idea - thanks very much for getting back to me and letting me know about this. I'm going to install the Deluge Docker container and set it up with Shadowsocks on my Streisand server. Hopefully it'll work as expected. Thanks again!

So I'm fairly new to OMV, linux and Docker but learning and it's been a steep but fairly enjoyable journey getting to grips with the Docker containers which are fantastic for adding functionality.

Not sure the etiquette or what is considered essential/obvious information for anyone to be able to help so please ask if I miss out crucial info about my setup. I'm trying (like many others apparently) to get Transmission to have it's data funnelled through a VPN cointainer (I'd prefer Wireguard if possible as I have set up and run a Streisand server on Amazon Web Services).

Ideally, I'd like to have a container per-service approach with a container for the VPN and any other services rather than an "all-in-one" approach which I've seen are available.

I've Googled this to death and my brain is melting feeling like I'm getting somewhere and then the final details are nowhere to be found so I really hope someone can help.

I've got https://hub.docker.com/r/activeeos/wireguard-docker installed with the name WIREGUARD

alongside https://hub.docker.com/r/linuxserver/transmission with the name TRANSMISSION

Transmission is set to use the WIREGUARD container for connectivity with the following "Extra arguments": --net=container:WIREGUARD and *seems* to work although without being able to access Transmission I can't test it.

The Wireguard container is configured as per the instructions here: https://github.com/activeeos/wireguard-docker

Now I tried and tried to get an Nginx container up and running as a proxy (tried both the official one and the linuxserver one) and modified the .conf file to try to be able to access the transmission web interface which is otherwise inaccessible while linked to the Wireguard container but I just couldn't get it working.

Searching Reddit I found a link to This page which supposedly details how to "Create a container behind the PIA VPN" which I thought I could use to create a docker network (as --link mentioned elsewhere in forums and guides is apparently deprecated) and access the Transmission webui.

However, that page includes commands such as "docker run ... --net=container:pia --name=myservice myservice" with the "..." clearly expecting me to know what goes there and I couldn't figure out how to fill in the blanks or put the pieces together.

I've reached a dead end with this and was hoping someone could provide a tutorial or list of step by step OMV based instructions using either the UI or Shellinabox explaining how to get Transmission to use Wireguard (or OpenVPN if i have to) and how to forward the Transmission Webui through either Nginx or using Docker network settings.

How do I set up a Docker network so that I can access the Transmission UI while it's using the Wireguard container? Or how do I set up an linuxserver/nginx container so that it allows me to access the Transmission UI?

I'm having almost the exact same issue/difficulty. I watched the TechnoDadLife video which was great in explaining the basic setup of the Environment variables it expects, but realised quickly that custom providers wasn't a simple thing to implement necessarily.

I'm using and want to stick with a custom Streisand server I have running on AWS which generates .ovpn files and it's these I'm trying to use with this.

In the Volumes and Bind mounts I've got:

Host path field: /sharedfolders/SSDSTORE/TROVPN (a shared folder in which I've put the .ovpn file from Streisand)
Container path field: /config

In the Extra arguments field I've tried: -v /config/x.x.x.x-direct.ovpn:/etc/openvpn/custom/default.ovpn (x.x.x.x is the ip of my server)

I've seen in the instructions that the script "adjustConfigs.sh" needs to be run to modify the .ovpn file with "your folder name as parameter" so that it works but I'm not sure either which folder it's referring to as "your folder name" (the folder with the .ovpn file in it or the ovpn folder within the Docker Dontainer?) or how to properly run the script.

Also, I'm not sure how to access the script itself to run it. This one would take some Googling if I was trying to figure it out on my own and I'm fairly new to Linux and the command line. I'll think i'll need to figure out how to access the adjustConfigs.sh file (In the docker container? or download a copy from the Gihub repo?) I have found a copy of the script from a branched project for Deluge but not sure it's the same or up to date?: https://github.com/hayduck/doc…/openvpn/adjustConfigs.sh

My trial and error approach would probably be to download the script file to a the shared folder with the .ovpn file in and run it guessing which path it expects.

Every time I run the image and create a container with the Environment variables and arguments, for whatever reason (misconfiguration of the paths or .ovpn file I think) the container seems to be constantly "restarting" and then disappears from the list of containers altogether as if it's deleted itself.

Any help with this would be greatly appreciated in how to set the paths correctly in the container and run the script to modify the .ovpn file. I'll keep trying stuff and if I figure it out I'll come back and update. Thanks!