testdasi

The B chipset does not have enough lanes for a M.2 + 6 SATA. If you read the manual (pdf can be downloaded from the manufacturer website), it will tell you e.g. if you use M.2 then 2 SATA ports will be deactivated. You need X chipset i.e. more expensive. To use the Ryzen 7, you will also need at least a low-end GPU as it doesn't have a integrated GPU.

I underlined the word "need". I would not imagine having a GPU entering low power mode is very high on the need scale.

Your point is not valid for the discussion in this topic. The only use at the moment for needing GPU driver baked into Unraid is for hardware transcoding. Even if one ignores the fact that pretty much all of the people requiring the feature are Nvidia users, the proportion of user base requiring hardware transcoding is certainly less than 80%. So no, not having this feature will definitely not drop the user base to arbitrary 1/5 that you mentioned.

2 of the 6 SATA ports are not through the chipset by 3rd party ASMedia ASM1061 controller. I don't tend to trust non-chipset SATA ports as I have had a few that didn't work properly. Also to run the 200GE on 3xx chipset, you will have to update BIOS, which may require an older gen CPU. So just to be safe get 4xx chipset.

Unraid does NOT decide which card to boot with. That is controlled by the BIOS. Most mobo BIOS will pick the 1st PCIe slot by default if there's a GPU there with no other option. So if you use a single GPU then by default Unraid will boot with that. If you pass it through to the VM then Unraid will lose the display (until the next reboot - there's no way to pass the GPU back to Unraid OS once passed through to a VM - but you can pass the GPU to a different VM).

Wrong place to ask for this. You need to talk to Linus Torvalds. All kernel control is by the power-that-be. Unraid is an end user of the Linux kernel.

I prefer to use the status file method e.g. Check if a status file exists. If yes, exit script. If files does not exist then touch status file (to create it) and then do stuff and once all done delete status file I find this method works really well with cross-script signalling.

I believe there are Minecraft server docker for Unraid (you can find the dockers / plugins in Community Apps). The general recommendation is if there's a docker then use docker. Only use VM as a last resort i.e. there's no docker available for Unraid. Diagnosing performance problems with VM can be quite a pain in the backside.

There's nothing weird about it. One of your backup has been running for more than 24 hours so it overlaps with the "once a day" schedule. You have to write your script properly to detect this scenario.

If you want Unraid to write syslog to a local file on the Unraid server then the remote syslog server line should be the IP address of that same Unraid server.

You very much certainly haven't or at least not paid sufficient attention to the details. He had a video explaining the pre-requisites for passthrough e.g. IOMMU. Watch his Unraid VMs playlist again please, particularly video number 6.

What did you put in the "Remote syslog server" line?

If the GT710 runs well with Windows VM then the problem lies with the Volumio OS then. Could be driver, settings, many things but I don't think it would be Unraid related.

It is a kernel problem which is outside of LT control. Kernel-related issues will never be fixed until it's fixed by the power-that-be (the elite few who control the Linux kernel codes) or someone creates a patch (that doesn't end up breaking other stuff) or Intel fixes it on their end (e.g. firmware update). The 660p is dirt cheap for a reason. QLC NVMe is like running a Ferrari with bicycle tires.

When you said "both cards test good", you meant they both work fine if passed through to a Windows VM? Which card does Unraid boot with?

The i7-6700k has iGPU that can be used for Plex transcoding. You should explore using that first before considering adding the GTX 1070.

Only Gigabyte allows picking which PCIe as initial display. Note: ONLY x16 length slots. x1 length slot is not pick-able. For the board you quoted, the GTX 950 is likely to occupy 2 slots so it will cover the x1 slot. You can only put the lowend GPU on the bottom most slot, which means you should get a single-slot width GPU. If your GTX 950 is the 2.5-slot width type then there's no where to put anything else. If you add the LSI then you have to use ATX motherboard (or try your luck with passing through single GPU).

The 1660 has 4 devices that all need to be passed through together for it to work. Your xml only has 2 of the 4. You need to vfio-ids stub the other 2 (or preferably all 4) - check devices 05:00.0 - 3 Also make sure you don't have any adblocker on the browser when doing Unraid settings. And do a disk check of your USB stick. SeaBIOS is fine but OVMF (i.e. UEFI boot) is better. It allows you to switch to Q35 machine type in the future should you desire (Q35 needs OVMF).

As always. Tools -> Diagnostics -> attach zip file.

Do you actually get processed killed because of out of memory error (check your syslog)? If not, that's just write being cached in RAM. Linux kernel will release RAM cache automatically as other things start to use memory. Remember: Unraid is Linux-based and with Linux, free RAM is wasted RAM.

You can try to limit docker memory usage by adding this to the "Extra Parameters" box (you need to turn on Advanced View) --memory=4G The above will limit the docker memory usage to 4G. That usually force the docker to behave less aggressively with memory usage. If the docker still misbehaves and uses more memory than it's allowed, it will be killed as a priority over other processes. Note that without the memory limit parameter, the process that is killed is not necessarily the culprit. It's usually just the one using the most memory.

Definitely NIC issue. Within the space of 1 minute, your NIC went down twice (and likely multiple times every time you run the rsync script). You can try restarting the server. That can help in some cases. Failing that, you need a new NIC (but hopefully a restart will be fine). Jan 31 23:18:46 Mediabox kernel: e1000e: eth0 NIC Link is Down ... Jan 31 23:19:05 Mediabox kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx ... Jan 31 23:19:09 Mediabox kernel: e1000e: eth0 NIC Link is Down ... Jan 31 23:19:13 Mediabox kernel: e1000e: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx

Is it a new patch or is it the same patch that was included and later removed from the RC because it broke other things that worked? The OP of the topic you quoted even share the procedure to compile your own kernel so perhaps it's best to compile your own kernel.

Pinning is not enough. You need to isolate. But first emphasis: you should NOT pin and/or isolate core 0 (and its hypertreading sister). Worrying about NUMA is going too far at the moment. You need to sort out the single-CPU case first. Have you tried assigning only 5 cores to the VM, leaving core 0 free? For USB sound, the best way is to pass through the whole controller. Leaving core 0 free may mitigate the issue to some extent (because the emulated USB controller has some spare CPU power to process).

I thought I replied to you but maybe forgot to post. What you are seeing is very typical of network issues e.g. random connection dropping. You should try to check connections on both end (or at least check the syslog to see if your connections are dropping). I have had overheating NIC randomly dropping off and back on in the past. Brute force is ok but if you don't resolve the connection issues, you may end up not being able to trust the content of the copy.