GRRRRRRR

let LLMs write my Resume, get new job, higher pay, put new HDDs get all new personal gear which I feel happy about at all times in all aspects feel sexy about my server Old HDD will have slower sectors so someone will need this HDD for their use case so I sell it cheap online. With information about SMART status screenshot of HDD Sentinel etc. Problem solved everyone very happy.

The OS can spin down disks, when they are unused for some time. The IOPS can be redirected to the nvme ssd cache, so that the disks will indeed be unused. The ReFS is configurable when to rebalance the stuff on the disks or at least should be. However I would advise against targetting this scenario for powersavings. For powersavings I will advise to go straight to huge SSDs. Unless you have some special situation then some special finetuning may be possible but still unpractical I would judge. For archiving I can advise UnRAID power savings are sufficient.

Yeah it's a funny business this secure file sync some corporations have it as their core business and product even. But you can do it on unRAID with OwnCloud container and WebDAV, or that sync-app available on all platforms (syncthing?), and have a cooler setup than the corporations, to share with your friends and family. 😁

Do like this: Do the easy thing "just do it" tm: use WoL to reduce powerusage and task-schedule all the stuff at night during the cheap electricity time. 1 system unRAID mostly default config as your stated requirements without GPU passthrough at first - can do WoL to reduce power Put the HDDs in the unraid pool and add the NVMe SSD as cache as described in a video I can provide if you need... 1 NAS Synology archive, can do WoL to reduce power 1 Gaming and Productivity workstation PC - turn On all power reduction tech however keep overclocking On to reduce overall latency 1x 10Gig connection for 10 usd this is a good deal however even 2.5gbps internet connection is just fine, if you truly need more than 2.5g you will know it without asking the forums. 1x system for experimentation of the advertised features and youtube inspirations - because youtube always encoures you to change modify your stuff, it takes a lot of effort to contionually change your setup so make it a least priority - stay stable longterm for productivity guaranteed and enjoyment without risking breaking the setup. This will give you the least grief and the most satisfaction probably because it will allow you to prioritize your actions without being dependent on your environment and tooling so much. Hope this helps to reduce the workloads and increase the efficiency.

This piece of rusted metal is hitting the target. 😈 Tightly packed. It's loaded with 12x 3.5" hard drives. Can stack up to 30 hard drives. Noise suppression. 4x 120mm active cooling. Dust suppression. Zero upfront cost, 60W min, 110W continuous, up to 280W peak power consumption. 1 gigabit fiber uplink. Automated content management per disk. Can't buy such a thing even if you tried. Made it in few weeks from spare parts. Used a few power tools to mod the chassis. No need to show you a pic, it looks like a PC on the outside, is a hellish monster on the inside.

Someone broke it. Wasn't me. Add static IP manually. Quote: README.md unraid_arc_g2 compiled unraid kernel images to get intel arc dg2 working this was compiled from https://github.com/thor2002ro/unraid_kernel (rc5) for me dhcp was broken un this RC so you may have to set a static ip if that happens to you as well changes: bzmodules & bzimages: CONFIG_INTEL_MEI=m CONFIG_INTEL_MEI_ME=m CONFIG_INTEL_MEI_TXE=m CONFIG_INTEL_MEI_GSC=m CONFIG_INTEL_MEI_HDCP=m CONFIG_INTEL_MEI_PXP=m CONFIG_INTEL_MEI_WDT=m bzfirmware: files from https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915 dg2_huc_gsc.bin dg2_dmc_ver2_08.bin /etc/rc.d/rc.local: line 129: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 132: /usr/local/sbin/plugin: No such file or directory /etc/rc.d/rc.local: line 132: logger: command not found /etc/rc.d/rc.local: line 138: logger: command not found /etc/rc.d/rc.local: line 147: logger: command not found /etc/rc.d/rc.local: line 148: frondos: command not found agetty: cannot open UNIX socket agetty: cannot open UNIX socket Welcome to Linux 6.2.0-rc7-thor-Unraid+ x86_64 (tty1) Tower login: up In a situation where you have command-line access but not network access, you can set a static IP address for your network interface directly from the command line. Here's a step-by-step guide on how you might go about doing this: 1. **Identify your network interface:** The first thing you need to do is figure out which network interface you want to configure. You can do this with the `ip link show` or `ifconfig` command. The output should list your network interfaces. Typically, your primary ethernet interface will be something like `eth0` or `ens33`. 2. **Set the IP address:** Once you know which interface you want to configure, you can use the `ip` command to set its IP address. The command to do this is as follows: ```bash ip addr add 192.0.2.1/24 dev eth0 ``` Replace `192.0.2.1/24` with your desired IP address and subnet mask, and `eth0` with your network interface. Note that you'll need root permissions to run this command. 3. **Set the default gateway:** You'll also need to set the default gateway for your network interface. You can do this with the `ip` command as well: ```bash ip route add default via 192.0.2.254 ``` Replace `192.0.2.254` with the IP address of your gateway (typically your router's IP address). 4. **Set the DNS server:** Finally, you'll need to set your DNS server. This is done in the `/etc/resolv.conf` file. You can edit this file with a text editor like `nano` or `vi`: ```bash echo "nameserver 8.8.8.8" > /etc/resolv.conf ``` Replace `8.8.8.8` with the IP address of your preferred DNS server. The above command uses Google's DNS server. Remember, these settings will be lost after a system reboot. To make these changes persistent across reboots, you'll need to edit the network configuration file for your specific Linux distribution. For Slackware, these settings are typically stored in the `/etc/rc.d/rc.inet1.conf` file. As always, be careful when making changes to system configuration files, and only do so if you're comfortable with what you're doing. If you're not sure, it's better to ask for help rather than risk making a mistake.

Productize it. 😎

I apologize for my outburst. I think it's a good project and the videos, and the interface, are quite good as well.

Hi, Do you have critical workloads and data on the unRAID or only non critical so far? 860 Evo 1TB sata3 - good for non critical media stuff such as running containers and VMs off of it, since it will probably not fail soon / in general I think this is better than the WD blue you mentioned (WDC_WDS100T2B0A) ., but haven't looked in depth. 960 PRO 512GB nvme - good for cache, vms and dockers if you feel like accelerating some things, will probably not fail soon. Keep an eye on smart and reduce writes if they become too much (maybe choose either burn it out for cache/ or for vms and dockers only) If you put them in current Desktop you can StorageSpaces and ReFS them together to get combined space and nvme speed without parity protection, to use for non critical fast storage such as Program Files and 'Media Scratch' if you are a creator ( I assume you use MS on the Desktop) Back that up to the unRAID in case something burst into flames (such as the filesystem versions with Windows OS different versions if you ever try to use older OS.) The second 1060 passthrough that could be used if ever need to have a guest use a PC for guest PC for example. A VM with a GPU can also be connected to a movie watching screen. The GPU also has digital audio output through the HDMI... Alternatively, the second GPU can reencode more streams if you share with friends, maybe with some TV stuff. It really depends how you really want to use your systems for your goals.

It could be there aren't enough PCI-e lanes, limited either by BIOS, CPU, chipset or physical electrical contacts, as well as there is also the possibility of power delivery creating the limiting factor. Try turn off some PCI stuff from BIOS/UEFI to see if the PCIe connection starts working. Didn't look inside the attached diagnostics, what is the motherboard and all PCIe chips and devices attached?

I will try to do it with IDE to cfcard but I need to reconfigure the cfcard from industrial 4MB to normal 1GB first, probably with some DOS tool or some SMtool or something. Any ideas how to de-industrialize the cfcard pleaze let me know. Cheers. Other than that the solution is to deploy the functions across 2x unRAID and stack some sync tech such as glusterfs.

But why with PCI-e switch? The x8/x8 should work like if you get a 4x nvme without PCI-e switch, the first and third nvme slots will work and the second and fourth will not work. The quicksync acceleration is when the Intel CPU has an iGPU and the two can work together to do some media stuff. If you put a CPU that has quicksync in a desktop system then it may work in this system as well.

The question is when VPN is not an option, right? When I absolutely must open the port to the Internet or an unverified system (what if the hackers are in the Plex itself already and get in trhough that?) in this case I put in OS level hardening as much as I can usually however it's a custom boundary based on the OS features + some security system features such as run as limited user, apparmor docker vm etc I turn on as many things as there are as well as do not put any actual data near the open port so it's like a DMZ ? with multiple checkpoints, nothing comes in that should not, and nothing comes out that should not. if it's http I put WAF. Infront I put a Suricata. And in front of that even a thirdparty scrubbing with a sensitive auto trigger, if I have to. And in front of that I can put a cloud CDN for additional protection and scrubbing. I can also put a weird analytical system that does weird anomaly pattern detection and even fine tune that algorithm for a very sensitive auto trigger. I can also put a few simple script that let me in separately and then my plex client that's with me can also go in. All that because some plex client devices such as SmartTVs at the GF house may not be able to VPN in.

In this case when the motherboard already has the bifurcation in place, and the PCI-e lanes available by the CPU?, you can configure that in BIOS to enable the bifurcation needed. In this case my suggestion is to populate the x16 PCIe- slot with an x16 to 4x nvme m.2 pci-e adapter, without chip on it. If it has a chip on it, it's probably a PLX/pci-e switch or a raid controller, which in this case is not needed and may not fit the power budget. However it is Low Profiles so this would be the product https://www.aliexpress.com/item/32991136466.html?algo_exp_id=81ea77fe-7586-426c-887a-731e559848f7-27&pdp_ext_f={"sku_id"%3A"66835179259"} unless you can find even cheaper somewhere or 4x nvme on both sides https://www.glotrends-store.com/river-p0080.html there are other options out there, Where do you prefer to ship from ?

My friends personally I would not buy the PLX etc PCIe expanders because of their high price mainly. I did it and I was not happy with the performance for the TCO. I was searching for when Linus said that sometimes you can use 4x nvme card in two of the slots and it works but didn't find it yet. Meanwhile I found a hardware modding guide which really peaked my interest however I do not really plan to apply it myself. https://linustechtips.com/topic/1463027-say-no-to-the-paywall-of-pci-e-bifurcation/ Did you say the HP has the bifurcation options? And why do you need "local bifurication" ?

On topic, don't fall for that trap to use a custom kernel that you will need to manually work on on updates and troubleshooting. Do the encoding on a separate machine with the kernel that will work. And even then, on MF topic, do not even do the encoding, with the MF GPU, or loose the benefits of the codec. What happened? Wanna argue about it? Put some facts on the table and we can talk. Otherwise, be comfortable in being destroyed in every which way you can think of arguing.

I have seen project fail and flourish and design and architectural ideas fail and flourish. Rest assured I kick your ass in the right direction I do not want to insult anyone in the way you are suggesting. However if I have gutpunched THE RESPONSIBLE person for a problem then I have done what I wanted to do. So take any hurt feeling to the responsible person for an exact problem. I am pointing out how things can be better actually addressing the root causes or exact decision points where the optimal solution is. The thing I said is a metaphor or analogy "a comparison between two things, typically for the purpose of explanation or clarification." You sir can not argue with me at all, I am sorry to say. The poster wants to encode and decode I pointed him and others in the EXACT precise right direction without any deviation whatsoever + I have kicked all your asses along the way. The kernel will be overwritten with each update unless the packaging system, if nay, clashes the versioning. If you are in the situation of having to ask this question the Operating system have failed you and you should backtrack to where the wrong decision was.

Well for one I see many of you are younger and susceptible to doing things in a trendy but less functional way, which will bite us all in the ass later, in terms of who knows what unforeseen edge cases and security swiss cheezing the entire system. This is why I am shouting at you about how it should be better. Such as not using JavaScript. Redirecting back to an unstable reasoning for doing things is not an answer to my fundamental questions to you. State the facts truthful as they are or be destroyed in seconds in any argument. I do not think that you or the function need my "constructive" observations. I think you all need you ass kicked back to the root of things for you to realize to do things properly. I am not saying you are a bad person, or have done anything very wrong. If there is not yet an easy to use optimized 'framework' for what you wanted to do then do it in the bloated framework, however when I ask you WHY TF you say the truth and no BS. The constructive you are looking for is - good job, Excellent job. But when you are done with the function, I still ask you uncomfortable questions. What kind of CI/CD security practices are going on with the community plugins anyway?

Do not encode the AV1 on the GPU. Do that on CPUs to get the space savings and preserve the quality. However, with this idea to decode for devices that don't support it, H266 is better. With messing with the Kernel you introduce additional risks to your unRAID storage system as well as dilemmas like the ones you are asking. Better to do your encoding on a separate system that can be reinstalled as needed, if you need to recompile the kernel, just to do the encoding. However please stop falling into old traps, these problems should have been solved decades ago. There are proper way of building things and improper way that will fall on your head with the first earthquake. Stop building improper and stop suffering.

And here I found some info how to compile kernel + modules for more specifically slackware maybe a module with be available that has this tty driver https://docs.slackware.com/howtos:slackware_admin:using_slackwares_kernel_build_scripts

I am not sure how that is supposed to work. But I remember there were some Ko files that were later integrated into the kernel itself. So something similar may be happening with this stuff. Meanwhile, you can try to compile this stuff https://github.com/torvalds/linux/tree/master/drivers/usb/serial to see what ko files come out of it if any... ?

What do you mean, added to the kernel? Shouldn't it be a ko file that can be insmodded when needed? I am confused. Please share more details. I use some stuff like this for home automation as well maybe it will be useful for me.

Well said well said. Always looking for ways to save 1000$+ /year on the electricity bills.

Short answer, use any network card 2.5G-100G it does not matter. (Use the Mikrotik switch with the 100G if you are wondering how to fit all that into a budget) Long answer, find one with lower power consumption which is also smaller, which can also be cooled adequately, which also has good Linux, Windows and BSD support. How long can you wait for shipping and how much money you can spend on that stuff? And why ?

My current build is combining many parts so I would use the Asmedia, after testing, to verify stability under full load. For new builds, if I spend new money, that investment goes into NVMe PCI-e for the newer chipset on the drive, lower latency higher IOPS If I have leftover cash and space inside the machine, I can put any cheap thing that performs stable after testing. If I need to swap specific drives often, I like Broadcom latest chipsets with support and firmware updates. If I don't want to buy that, I like the latest AMD sata controllers.