tjb_altf4

Swapfile is only needed in a few specific use cases, and shouldn't be needed in most scenarios. Unraid has no swap by default, this plugin adds it for those users that need it.

PCIe versions aren't too bad, you just need to power via a riser, same as the ones GPU miners use. Some non-pcie pwoered options I know: Intel RES2CV240 (regular molex power connector) Dell 5R10N (needs custom pinned 4 pin connector)

Its a Samsung hardware problem, you can either add the kernel arg to bandaid fix the problem, or update the nvme's firmware to fix at the source. Both solutions are detailed in the linked thread.

If 1m from the motherboard/HBA isn't enough, you can run the longer SAS lengths between the HBA and a SAS expander (that you'll need to add). From the expander you are then again limited to 1m for SATA. In this scenario, make sure parity HDD and any SSDs connect directly to motherboard if possible, to ensure they get maximum bandwidth.

If your problem is incorrect temp reporting, which is specific to the newer samsung nvme devices like the 980, this thread from 6 months ago has a fix. But are you sure you don't simply need to change the temp warning ranges for your nvme? The default temp ranges are 45-55C for all storage devices in Unraid, which is too low for nvme and will definitely trigger warnings when under a reasonable load. I usually change this for my nvme devices to 55-65C (which is still a bit below spec for most nvmes) and rarely get temp warnings. This change can be done by clicking on the device name (e.g. Cache, Cache 2), from the MAIN tab

https://emby.media/community/index.php?/blogs/entry/464-emby-for-samsung-tizen-tv-now-available-for-usb-installation/ Requires using global app store, or monthly usb re-install. Tizen app is nice, but support is not great outside US/UK region Google Translate Erfordert die Verwendung des globalen App Store oder eine monatliche USB-Neuinstallation. Die Tizen-App ist nett, aber der Support ist außerhalb der USA/Großbritanniens nicht großartig

Simple answer is you are killing any running process still accessing files in your user shares.

Reads like its host managed SMR, which has some support in recent Linux kernels. @JorgeB would likely know if its possible on Unraid.

Quick one to check, did you plug the QSFP cable into the left side (square icon) of the IOM6? Right side is only used for daisy chaining (or secondary path).

If you have those drives (and only those drives) for the VM on a HBA which you then bind to vfio-driver, what you want to do would work. However, the next release of Unraid is under development and ZFS has been mentioned to be officially supported for this release, and already is unofficially supported via community plugin... so I would consider springing for the Pro licence and running ZFS on Unraid itself.

Don't forget to also use different appdata paths for each instance (in addition to the previously mentioned unique name and port).

should work, I'm running on the same spec hardware and cpu on unraid already

All PDFs are working for me

"This year we put a 12 on the box"

Not necessarily the fix, but you are on a legacy certificate (using the old url schema) that will expire in a few weeks, and should be updated ASAP.

Should be a sff-8087 to sff-8087 cable, but do check your own card's connectors. You can run a single cable, but if you run two (from a single HBA only) you'll increase the overall bandwidth available to the expander. Expander uses pcie for power only (no data). To keep those slots free for other devices you can use any slot including a 1x slot, or none at all and use molex power connection instead. HBA + Expander will have less bandwidth than 2x HBA, but that is a limit you probably won't notice unless you are maxing out drive connections. A forum member JorgeB has a great benchmarking thread which can give you an idea on speeds to expect. Typically the best performance scenario is to put parity and any sata ssds on the motherboard sata ports, so these will always run at maximum capacity.

You still need one HBA (e.g. H310) connected to the expander. for the expander to work. The expander allows you to run more disks off the HBA by sharing the bandwidth.

I noticed if you are navigating paths via unraid file manager plugin, then try to use GUI search it redirects to a malformed URL Normal operation When in file manager paths or in some cases shows like this

Remove the IP assignment for the vlan in network settings.

The vgpu_unlock is one thing... I understand devs don't want to poke the bear (Nvidia's legal department) giving consumer hardware access to enterprise features, although there are legitimate uses cases such as optimizing vgpu profiles. But vgpu support where it is fully supported and within Nvidia's licensing would be nice, I do have a Tesla P4 for this purpose which I will be using with proxmox on my new server setup, but I'd be happy to help with testing in Unraid should an official plugin be considered for development

Just to check the basics here, there should be at least one path mapped into the container that corresponds to the drive you have setup

You need the Advanced View toggled on (top right) when in the template view for your container. The setting will be visible in all path configs.

Unraid GUI > Tools > New Permissions is a once off tool to clean up perms, it also shows the reference permissions/ownership that are set. (New Permissions tool)...changes file and directory ownership to nobody/users (i.e., uid/gid to 99/100), and sets permissions as follows: For directories: drwxrwxrwx For read/write files: -rw-rw-rw- For readonly files: -r--r--r--

If you use the Slave option, the path should show in ZoneMinder

I just reread your first post, I think you talked about 2 seperate things. You stated the disk is mounted here: "/mnt/disks/212307B02", that reads like it has been mounted using UD. In which case if you want to use that option, the docker path being passed needs the slave option to be used. EDIT: Hoopster just beat me lol