Beiträge von mischka

Such a computer does not only consist of the processor. What is important is how well the individual components are integrated, how good is the Efi, how well are the C-States implemented, etc. etc. etc. And that's where I have my doubts about these Chinese. How good is the passive cooling in the Minix? Which kernel are you using? For the N100 you should use kernel 6.2 or higher (see kernel plugin from OMV or Debian backports. The N100 on my Asus board also runs at 1% idle (kernel 6.5.x) under OMV. The Optiplex is much older than the Minix. Dell (just like Lenovo, Fujitsu and HP) knows its craft. The Optiplex will not consume more power than the Minix in idle, if everything is set correctly in the EFI, despite the old i5-6400T. These old mini PCs from the well-known companies are usually very well optimized.

The certification for 80 plus up to and including platinium is window dressing and only starts at 20% load, i.e. at 10% the efficiency, no matter how good or bad, does not affect the certification. 80 Plus Titanium starts certification at a load of 10%. The efficiency must then be at least 90%. With the SFX600 power supply, the efficiency will be worse at a load below 10% than with the SFX450 at the same load. Most of the time, a NAS works at idle. With a MiniPC Fujitsu Esprimo Q556/2 with i5-7500T with 8GB RAM, a 32 GB and a 4TB SSD, the power consumption in idle is about 8 - 9 watts. This MiniPC has a 60Watt power supply with very high efficiency... even at low power consumption. With your components, you will hardly get below 20 watts. Especially since the Asus board is not particularly economical in idle despite the N100. Despite Pico PSU (120 watts max), I didn't get a power consumption below 12 watts. Was usually at 14 watts in idle.

Does video hardware acceleration work on Debian12/OMV? This is a basic prerequisite for passing them on to the container. What is the output of vainfo under ssh in the terminal? Intel's 12th CPU generation with integrated GPU (e.g. N100) requires at least kernel 6.2 for the correct function of video hardware acceleration (see Soma for kernel version control). I don't know how it is with the 11th generation. The following packages must be installed: intel-media-va-driver (for hardware-accelerated decoding... which is usually enough) or intel-media-va-driver-nonfree (if hardware-accelerated encoding is also to be used) and vainfo to control the function of the video hardware acceleration.

Millions of Qnap and Synology with ext4 and without ECC are in use. Very few users of these NAS systems complain about data loss. You can insure yourself against any risk. Whether this makes sense and the costs are justified is something everyone has to know for themselves. I'm happy with ext4 and multiple tiered backups.

All with a translation program. When it comes to network speed, the device with the lowest network speed always decides. If the NAS has a 2.5 Gbit LAN and the switch can only do 100 Mbit, the network speed will never be higher than 100 Mbit/sec (about 12 MByte/sec). Usually, an N97 or an N100 is enough for most tasks. TDP is less interesting because a NAS usually runs in idle. Even a Fujitsu Esprimo Q556/2 with an i5-7500T consumes less than 10 watts in idle mode with 2 SSD. It depends on how well the manufacturers of the hardware have optimized this and also how high the efficiency of the power supply is in the low-load range. The 65 watt power supply of the Fujitsu has a very high efficiency. A normal PC power supply (also 80 Plus) is less effective in the low-load environment.

I also own the Asus board. This disappoints all along the line. With 2 SSD and a 2.5 Gbli LAN card with 100Watt Pico PSU, this consumes about 14 watts in idle. The 2.5 Gbit Lan adapter that is plugged into the PCIex slot strangely only works when the inner 1 Gbit Lan is switched on. If the speaker is on the board, it complains loudly that no monitor is connected (dummy provides a remedy). WOL works via the PCIex 2.5 Gbit LAN card. I have also not been able to turn off Secure Boot in the UEFI. In addition, I could not activate CSM. And there are a few other things that bother me.

In the end, I run OMV on a used Fujitsu Q556/2 with i5-7500T for little money (leasing returns) with a built-in 2.5 Gbit network card in the Wifi slot (m.2 A+E key). Full performance, power consumption in idle very low and this is the only PC/motherboard of my tested PC/motherboard where WOL works with the built-in 2.5 Gbit LAN card in the Wifi m.2 slot.

New RAM is integrated by the computer's BIOS/UEFI when the computer is started and has less to do with OMV/operating system. If it is not visible after installing and starting the computer for the first time, then it is incompatible or defective or... Maybe put the Ram on other slots of the board or swap the slots of the Ram bars with each other. This sometimes helps, especially with older computers.

I don't understand the point of running Proxmox with a VM running HomeAssistant (?). The J4005 processor with its two cores is quite weak. If you only need HA, that would make more sense without running Proxmox. I wouldn't expect the CPU to add another OMV VM to the HA VM in Proxmox. If OMV and HA are to run, I would install OMV natively and run HA in OMV as Docker. The whole thing without Proxmox.

If OMV can no longer be started at all and you cannot access it with ssh, connect the OMV PC to the monitor and keyboard/mouse, start with a USB stick with a Linux live system and check the OMV partition with fsck. The easiest way would be to download a gparted-live-iso. Then you can work on a graphical interface. With Rufus or balenaEtcher you can create a bootable USB stick on another PC. Then start the OMV PC with this USB stick, then run gparted. In gparted, unmount the partition with OMV, check it and have it repaired. If you're lucky, your system will work again afterwards.

Google search "gparted-live-iso".

Ab Setting up NFS Client:

How to Configure NFS Server and Client on Linux Mint 20 – LinuxWays

I don't understand the whole thing. You can also transfer the data volumes between Windows and OMV at night. After that, OMV can be shut down automatically via autoshutdown. That will probably work with Windows as well. I only use Windows for a few programs. I installed Proxmox for testing. Among others, Windows 10 and OMV each run in a VM. If you use Proxmox to virtualize Windows 11, it should be at least one CPU's with 6 cores and more. You also need enough memory. Windows 10 runs well for me under Proxmox on an i5-9500T with 32 GB RAM. I haven't virtualized Windows 11. That just annoys me too much.

Many 2.5-inch hard drives with enclosures, such as Toshiba Canvio Basic, go into hibernation mode after a certain period of inactivity. This is implemented in the firmware of the controller of these hard drives. These plates have difficulty waking up when requested by the Raspi. A script that keeps the disks awake helps. Of course, this is at the expense of electricity consumption. As for the power supply... it is important to use the original RaspberryPi power supply. This is usually enough for a disk on the USB3 and an SDCard. Never use any chargers or power adapters from other vendors. The Raspi 4 needs the original power supply with 15.3 watts. The Raspi5 does that with 27 watts.

The only thing that needs to be deleted or commented out is the first line in your red flag UUID=... The second line mounts the DVD drive (/dev/sr0). Everything else is described in detail in post #13. I would do without the swap partition altogether. It is neither active nor integrated as they have created it. Allocate enough RAM to the OMV VM in Proxmox (8GB). Then boot up again with gparted in the OMV VM, delete the swap partition and use the space to enlarge the OMV system partion. That's it from me!!

No one knows how you went about enlarging the system partition of OMV!. Did you delete or move the swap partition in the process? As long as nothing concrete is known about it, nothing more will come from me!

If Jellyfin is running in an unprivileged LXC, it can only access network shares from the OMV VM in a roundabout way. The LXC Jellyfin container must be privileged. But even then, there are enough problems. What works very well is to run Jellyfin as a Docker container in the OMV VM. To do this, you need to install the Compose plugin in OMV. For me, OMV runs very well in a Proxmox VM with Jellyfin Docker container. Access to OMV's data is then no problem for Jellyfin.

I'm guessing that when you enlarged the OMV system partition, you deleted or moved the swap partition?! You can't just delete or move a swap partition. It must be deregistered in advance. This also includes deleting or commenting out the entry in the fstab. Remove Swap Permanently In the OMV VM via ssh:

1. Using the swapoff utility:

sudo swapoff —all (or sudo swapoff –a)

2. Comment out or delete UUID of the swap partition in the fstab

sudo nano /etc/fstab

3. Comment out the UUID of the swap partition in the following file:

sudo nano /etc/initramfs-tools/conf.d/resume

4. Execute these commands at the end (both in this and in the in opposite order):

sudo update-grub

sudo update-initramfs -u

To enlarge the file system, you have to do this in the OMV VM. In Proxmox itself, you only increase the size of the virtual drive. The best thing to do is to integrate a gparted Live ISO into the OMV VM, start from gparted and then enlarge the OMV file system or system partition. If there is a swap partition directly behind the OMV system partition, you have to deactivate it first (preferably via ssh in OMV), then delete it, enlarge the OMV partition and then create and mount a swap partition again if desired. How to do this by disabling the swap partition... see Google.

I had the same behavior on several Asrock m-ITX boards with J4105, J4125 and J5040 after the boards were de-energized. There is nothing in the syslog because Linux is not even started. My mini PC that I use now (Fujitsu Esprimo Q556/2) does not have this behavior. I never found out why that was the case with the Asrock boards.

It is best to install kernel 6.5.0-0 (backports). Then the RTL8125 works without additional driver installation (at least for me). This should probably be implemented from kernel 6.2.0 onwards.

I don't understand that either. Apart from higher power consumption, this does not bring any advantages. From Skylake (UHD 530) onwards, the hardware acceleration transcoding works very well with the CPU-Graphic. There is no need for an additional energy guzzler. Then there's a Seasonic FOCUS GX-850. An oversized Plus 80 Gold power supply with 850 watts, which has an efficiency of 90% with a power requirement of 20%. Energy still seems to be very cheap in some countries.