Possible limitations of OMV when using command line ?

The basic rule of thumb is that anything that is configured in the OMV interface should not be messed with in the cli. If you mess with a setting via the cli that omv controls, you can break omv or the setting will be overwritten the next time a chage is applied in the onv gui.

Some thing's like samba have "extra options" fields that you can use for customizations, but there are also some customizations that can be done via omv environment variables. If the environment variable exists, it's best to use them instead of the extra options fields if you want the customization to apply everywhere.

Other than that, the developers can speak in more depth about the "under the hood" workings if there are more questions.

Well you only mentioned samba and rsync and as I said, samba has customization capabilities. rsync is rsync. it has a gui page, but you can also do it via cli if you wish. You can make custom scripts and run them however you see fit, or schedule them (or normal cli commands like rsync) with the task scheduler (basically cron with a gui).

Coming from proprietary solutions, I think you will be surprised with the flexibility of OMV, just don't expect an app store like those other ones. Docker is basically your app store. However, without more direct questions I have no idea on what you are trying to find information on.

As an example of what I mean by flexibility, my system runs docker with 50+ containers, kvm with several qemu VM's and several lxc's. Some of those things are internet exposed through a reverse proxy. I have custom scripts made to do automated rotating backups for all the dockers, vm's, OS, and important data the way I like to do it, even though there are backup options in the plugins. I use portainer to manage my dockers, even though there is a docker compose plugin because I find it suits my needs better. Most native debian cli tools can be added if required, but be wary of ones that would cause updates to php or python. The omv interface uses nginx, so you can't install apache natively, but you can use a docker for it or another nginx to allow you host pages or do custom reverse proxy setups (there are also nginx-proxy-manager, SWAG, traefik, haproxy, etc. dockers that are prebuilt reverse proxies). Things like webmin and olivetin can be installed natively as the web servers they use are "micro servers", not nginx and they use non-standard ports so there is no interference.

Once again, this is my setup as an example. It's the best I can tell you without the afore mentioned "more direct questions"

***edit***

just a point to extend what Aaron said about webmin. You can install it (I have it installed strictly for the easy file manager), but it will not allow you to circumvent the omv configuration management he mentioned. OMV is still the "final word" on configs. Installing it without OMV would give you full unrestricted control the same as if you were using a GUI-less server.

Quote from Quacksalber

After some sleep, I realized that my answer yesterday night was rather impolite. Please accept my apologies.

No worries. No offense taken.

Regarding your other points:

I don't use NUT, so I can't comment

Any rsync stuff I do is not using an rsync server

I have custom scripts I run that do a similar job to rnasnapshot using rsync, so once again I can't really coment.

I don't use BTRFS. I have a mdadm raid 5 and a raid 1 using xfs, all with the same size drives so I am not trying to mix and match. If you are trying that, the best route with OMV is to use mergerfs with snapraid on top, designating the largest drive as the parity drive. Most prople doing it are probably using ext4.

Quote from Quacksalber

I do have two questions, that I was not able to find out during my experiments with the virtual machine:

How do I know, when / if a hard disk fails?

I would have expected some indication on the WebUI, that I can't miss.

But I didn't see anything obvious...

The other question is about backing up the system / installation.

I will have the system / installation on an external USB stick. I know, that there is a plugin, that is meant to backup that USB stick to another one, but I couldn't figure out (yet), which of the many plugins to use for that.

Display More

Disk failure would probably be best looked after using the monitoring and notifications settings. SMART monitoring can warn you before a disk completely fails.. This needs to be enabled per disk. There is SMART widget you can put on the dashboard. There are also notifications that can be configured to send you an email alert when something fails, and of course in the case of a drive, the SMART widget should also reflect the failed status.

If you are installing on a USB stick, make sure you install the flash memory plugin. It will redirect the logging to RAM and only flush it to disk occasionally. This will greatly reduce writes to the stick and increase the lifespan of a USB stick OS drive from months to years. As for backups with a USB stick install, I would recommend keeping a clone of the usb stick that you freshen up periodically just in case that USB goes read only as they have a finite number of write cycles. Clonezilla or dd can look after the cloning.

As for the actual data backups, I personally use fsarchiver to backup my ssd os drive, as it can back up a live filesystem, but as with a lot of my backups, this is a custom script I run to do things the way I like to do them. You can however use the backup plugin if you want to use the built in tools. This plugin can use fsarchiver, dd, borg, or rsync to backup the drive.

You should also think about installing omv-regen. This is a brilliant little cli application written by Chente, that will gather and backup all the settings that you need to restore all your configurations on a freshly installed system, and it can even update the system before it does the settings dump. This is great for ultimate disaster recovery or even just to check the configs and replicate a section if you happen to "nuke a bunch of settings and can't remember or figure out what they were before.

My backup solutions is to use fsarchiver doing rotating full filesystem backups, and omv-regen doing the rotating settings dumps as an extra safety.

Quote from Quacksalber

hat the recommended way is, to have a second stick in a second USB port, so you can just change boot order, in case the old USB stick fails. This second USB stick must be regularly (e.g. once a week) updated with a clone.

No.

You can't have 2= OS drives connected at the same time.

Quote from Quacksalber

There would not even be the need to mount it, since cloning works directly on the block device.

Your BIOS will see 2x boot drives with the exact same UUID.

It doesn't matter if it's mounted or not.

Quote from Quacksalber

omv-regen really sounds nice. Haven't read about it yet. Thanks for the hint. Where do I get it? apt-get?

I will definitely install the flash memory plugin. This is already on my list.

And yes, I want to keep a clone of my installation on a second USB stick. I have read here in the forum, that the recommended way is, to have a second stick in a second USB port, so you can just change boot order, in case the old USB stick fails. This second USB stick must be regularly (e.g. once a week) updated with a clone. There seems to exist a plugin specifically designed for it, but I wasn't yet able to identify this. I will look into fsarchiver. Maybe its this.

As to the notifications: During my experiments I didn't activate neither monitoring nor notifications. Didn't make any sense for me in an experimental virtual machine. So yes, this might be the reason, why I didn't see anything...

Display More

The best thing to do to clone the drives would be to boot into clonezilla and do a disk to disk clone (or if you don't want to keep 2 around at least do a disk to image that can be written back to another usb stick) or to use dd to do a disk to disk clone.

omv-regen can be found in the guides section of the forum:

Quacksalber

Since you edited your post, I need to add more info to mine:

You CAN'T have 2 clones attached at the same time on ANY system.

Same goes if you clone DATA disks.

If they are boot systems (the OS drive of OMV, for eg), your BIOS won't care if it's a Windows, Linux, Android, whatever... boot drive.

The BIOS will only try to launch a bootable UEFI/MBR drive and it will do it on the 1st one it finds. (Not always the same)

Since it is a 1:1 copy of another, you have no idea which one you are booting. Same goes with the system.

What is even worst is, you can end with the system booting from one drive and mounting the rootfs of the other, because it has the exact same UUID. You will never have a clue that you booted with the USB stick on port1 and the system was mounted with the rootfs of port2.

Bottom line on the idea of cloning the USB stick is:
Either
Power down, clone to another, use the "another"

OR

Make a dd-full to img of the OS drive

Flash to "another"

Use "another"

Nowhere, you'll see to have BOTH connected at the same time

Quote from Quacksalber

Does this UUID belong to the hardware, the partition, the file system?

glossary - What is a UUID? - Stack Overflow

The UUID in Linux operating system explained | Linux Today

Quote

It's an identification number that will uniquely identify something. The idea being that id number will be universally unique. Thus, no two things should have the same uuid. In fact, if you were to generate 10 trillion uuids, there would be something along the lines of a .00000006 chance of two uuids being the same.

linux what is this UUID thing at DuckDuckGo

It can be a FS created, a shared folder, a pool, etc...

Debian/omv gives each filesystem a uuid and the mounts are based on uuid because the device identifiers are dynamic (ie. a drive identified as /dev/sda on this boot might be /dev/sdb on the next one depending on the order the os registers them). Duplicate uuid, can therefore “confuse” the os as soma mentioned.