TechTalk – Tagged "RAID"– Page 2 – TechMikeNY




flag How To flag Media Server flag Plex flag Storage flag tips and tricks flag TrueNAS flag Tutorial flag Virtualization flag Anatomy of a Server flag CPU flag Gaming
See more...


Hey Aaron,

To get your GPU recognized for transcoding, you’ll need to select the GPU in TrueNAS and turn on hardware transcoding in Plex. To do that:

1. Select Plex from your list of TrueNAS apps and edit its config. This will look different in different versions of TrueNAS.
1A. If you’re using “Cobia”, clicking the “Apps” button on the left side will pull up a list. Find Plex on the list, look in the box labeled “Application Info”, and select “Edit”.
1B. If you’re on “Bluefin”, the “Apps” button will take you to a screen of apps. Once you find Plex, click the 3 dots in the top right corner its little box, and select “Edit”.    

2. Scroll down until you see “GPU Configuration”.
3. If the section doesn’t have the “Select GPU” dropdowns, you’ll need to click “Add”.
4. Click the drop down to select the type of GPU you’re using. Select “Allocate 1 xxxxxxx/xxx GPU”. (xxxxxxx/xxx being your brand of GPU.)
5. Once you’ve selected your GPU, click “Save”, and wait for Plex to deploy.
6. After Plex has deployed, hop into Plex settings in the WebUI and select “Transcoder”.
7. From the Transcoder settings, find the checkbox that says “Use hardware acceleration when available”. Check that box to turn on hardware transcoding. (The checkbox will be the setting directly below the “Transcoder quality” dropdown.) 8. Remember to hit “Save” to apply the change.

And you should be all set. Once you configure your GPU in TrueNAS and turn on hardware transcoding in Plex, you’ll see “(hw)” written at the end of the stream data.

If you’re using a Nvidia GPU for transcoding, you can see if it’s being utilized by Plex when you have a stream playing that’s being transcoded. To do that, you can use the “nvidia-smi” command in the TrueNAS shell. (More on that here:

Hope that helps!

— TechMike

I’m struggling with this type of setup on Truenas Scale as I’m trying to get a GPU recognised for transcoding. I’m not seeing any provision for this in the setup. Have you been able to pull this off?

— Aaron

Thanks for sharing! Yes, the fix in that Reddit thread will work; as we mentioned in Step 11, you’ll need to make sure that the top fields are correctly filled out when you go to disable host path safety checks in “Advanced Settings”, aka “Kubernetes Settings”. (This also includes selecting “” from the “Node IP” dropdown.) Otherwise, if your port forwarding isn’t set up properly, Plex will not be able to see the server remotely.

— TechMike

In order to see my server from and also use the mobile apps, I had to change the settings in Plex and the apps settings in Truenas apps (interface and gateway fields) like explained here:

— Michel

Hey Logan, 

Thanks for your question. Start by making sure that “Plex Data Volume” and “Plex Config Volume” settings for Plex are pointing to the correct datasets (see step 13). If that’s all set properly, you may be running into one of a few things. 

We found that in order for Plex to recognize our media, we had to create subfolders for each individual piece of media and give the subfolders whatever name we wanted the media to appear as in Plex. 

If it’s a folder structure issue, here’s what you can do: Let’s say you want to add Jaws to Plex, for example. Your structure would need to be Media > Movies > Jaws > Jaws.mkv. If you want it to show up as, say, “Scary Shark Movie”, your structure could be Media > Movies > Scary Shark Movie > Jaws.mkv. (Note that changing the folder name will also change what Plex searches for to populate the movie info/poster.) 

Now, if that doesn’t solve the problem, then it may have to do with the permissions for your Media dataset. To check what your permissions are set to, go into the “Datasets” menu on the left, select your Media dataset, and click “Edit” on the permissions box. First, double check that Owner is set to “apps” for both User and Group. Then, make sure that the “Read”, “Write”, and “Execute” permission boxes are checked for both User and Group. 

Let us know if you need help with anything else!

— TechMike

Hey TechMike,
Thank you for the detailed guide, I made it through to the last step without issue, but seem to be having problems adding media to my server. Like in the other comment you responded to, I can’t get plex to recognize my media whatever I do.

I tried separating the files out into folders inside of Media called Movies and TV Shows, but Plex won’t seem to recognize anything past one folder deep.

It also doesn’t add the media (even if it shows up when I click the Data Tab when selecting a library) if they are in the root of the SMB Media folder.

Any help is appreciated.

— Logan

Excellent guide! Everything went very well with my first foray into using TrueNAS scale using this. I had used OMV with docker and Portainer before and this is even easier.

— Josh

Hey Trey,

Thanks for your question. In the final step of adding a Library to Plex, selecting “/data” during the last step is not actually adding data to Plex; it is directing Plex to look for media within the location, “/data”.

Backtracking to Step 8, we set up 2 datasets; 1. “Plex Config”, for Plex’s configuration files, and 2.“Media”, to store your videos, pictures, and any other content that you want Plex to access.

How does content end up in “Media”? That’s where the SMB share comes into play. The purpose of the SMB share is to be able to easily transfer your media files from your computer and into the “Media” dataset. This dataset is what Plex will access through the designated folder called “data.”

During Step 13, when configuring Plex’s “Storage” settings, you need to check the “Enable host path for Plex Data Volume” and set the host path to the “Media” dataset. By doing this, you’re essentially telling TrueNAS to link the “Media” dataset to the “/data” folder.

Then, when you go to add a Library and hit “Browse Media Dataset”, you’ll need to select the “/data” folder. You’ve set everything up so that whatever is in “Media” ends up in “/data”.

Hope that helps!

— TechMike

Last step to add Library to Plex doesnt work. If media is the directory and shared via SMB then why you adding data to Plex.

— Trey

Hi Danillo,

Thanks for your question. If the server does not boot into Windows after installation, you’ll want to check the boot type in BIOS – it should be set to UEFI. You may also need to enable PCI bifurcation in the BIOS as well.

— TechMike

Good afternoon, after the whole process, I did the installation
of Windows Server 2019, but after installation the Server does not boot into Windows.

— Danilo

Hi Dave,

Thanks for your question. The R640 does in fact support both PCIe SSDs and PCIe Bifurcation. However, it is dependent on the operating system containing the correct drivers for the PCI card or PCI SSD adapter. 

We cannot guarantee any part which we cannot test for compatibility, and we haven’t tested these servers with the 3rd party PCI NVME cards or 2.5" NVME adapters. Installing one of these cards may cause the server to run in safemode, spinning the fans up to 100%. Usually, the 2.5" NVME adapters are not compatible with the SAS/RAID controllers used in these model servers.

— TechMike

What do you think of those NVMe M.2 adapters which plug into a PCIe slot? You can find many inexpensive PCIe adapters on eBay which support 1, 2 or 4 NVMe M.2 drives. The ones which support 2 or 4 NVMe drives require a motherboard which supports PCIe bifurcation, but the R640 apparently does. Here are a couple of search strings you can use on eBay:

PCIe M.2 NVME profile -SATA

PCIe M.2 NVME “supports 4” -SATA

Alternately, if you want to go the U.2 route, there are M.2 NVMe to U.2 enclosures for under $10. What do you think of them? Here’s an eBay search string:

M.2 NVMe PCIe to U.2 2.5 SSD

— Dave Burton

To the webmaster, Your posts are always informative and well-explained.

— Joellen Snoddy

Thanks for the question, Dean. It’s important for NIC’s to support more than just the fastest speed to ensure compatibility with older connections and cabling; network environments are diverse, and different devices or network segments may have varying speed capabilities or requirements.

As for choosing the right NIC, your best bet is really just thoroughly researching your system’s requirements. The connection should match with your existing network hardware; however, the components on the servers are easily upgradeable, so remember, you’re not stuck with one type of connection forever. You can add a PCI NIC with a different type and speed of connection and you can change your network switch to match the new connections, be it 10GB Copper/RJ-45 Connections or 10/25/40GB Fiber SFP connections.

— TechMike

Why is it important for NICs to support more than just the fastest speed, and how do you choose between which NIC you might need?

— Dean Moore

Hey Mike,

Great question, thanks for asking! Dell doesn’t have a central directory with that information, but we compiled a table together for Dell’s NDC’s compatible with 12th and 13th/14th gen servers.

Here’s the link to copy and paste if you’d like to check it out:

— TechMike

One thing that is daunting is the expansive list of NICs& NDCs have boards that look equivalent (eg 4×1G RJ45) but have very different prices and only Dell bingo numbers to distinguish them. given the evolution of the Ethernet chips, knowing which part is on the card is golden. Is there an xref somewhere for the Dell novitiates?

— Mike O'Dell

Hey Zach,

Good catch! Just fixed it, thank you for letting us know.

— TechMike

Good info, thanks. Under “Number of ports”: You have the captions and the cards mixed up, which is a bit unfortunate as the caption is part of the image!

— Zach

These are wonderful ideas to keep old tech in our lives in a creative way!

— Jeff

Hey George and Adam,

Great suggestions! Thank you for letting us know what you’re interested in learning about. We incorporated info & recommended configs for media servers, routers/security gateways, gaming servers, administrative servers, and both types of hypervisors into our part 2 of this article.

Here it is if you’d like to check it out:

— TechMike

What about virtual machine hypervisors like Hyper-V or VMware ESXi servers? You should include that in your next article. :-)

— George Lancina

So, you covered mostly server types no one is likely to run at home and ignored media servers, routers/security gateways, gaming servers, administrative servers (Ubiquiti admin, etc), and general-purpose virtualization hosts/hypervisors (of either VM or container variety)?

— Anonymous

Hi Dustin, Yes, that is correct! That is the primary distinction between these two servers, but they both have the flex bay option available.

— TechMike

Thanks for sharing the images on what needs to be purchased. When you mention 12 bay vs 24 bay for the 720xd and 730xd this distinction is whether your server takes large form factor (LFF 3.5 inch) or small form factor (SFF 2.5 inch) hard drives right?

— Dustin

Thanks, Chad! We’re so happy to hear that. (And we’ve got some more in the works!☺)

— TechMike

Hello, Great videos! I find your videos of value and use.

— Chad

Don’t forget that some (all?) machines will not POST with a mix of LR-dimms and R-dimms! I have an r730 and an extra 256GB of ram that does not make it happy.

I just have it here, on my desk. Taunting me.

— Christopher Green

Hi Craig,
Great question. This probably will not cause any issues with virtualization, as ultimately, the CPUs will use RAM that hasn’t been allocated yet. Admins decide how much of the CPU and memory to use, but the server should have no trouble selecting where to use it from.

— TechMike

we have several Dell r430 of which we have purchased there, this is a general question about the R430, my question revolves around the fact the ram per cpu is different CPU1 has 8 dimms and CPU2 has 4 dimms, we are populating all slots with 32g chips so does this create some sort of issue in a virtualized environment considering how much ram per cpu as assigned to VM’s.

— craig sutphin