Vitalsignser

Since my MB has two network ports and my SFP28 card has two ports, if I create a third subnet with a direct connection between the VM NIC and MB NIC (ie a third network connection) then I should be able to share across this? I would need to make sure that the VM cannot get internet access from this 3rd connection. In Unraid - can a two port card be setup with two different subnets (ie same iommu with two separate subnet addresses)

I'm about to setup a Windows 10 (or 11 haven't decided) VM. I have two spare Sata SSDs. I was going to install the VM on one instead of a vdisk and use the second as a data storage disc for the VM. The VM will also have a separate passed through NIC on a different subnet which goes to a dedicated VPN port on my router (goal is to prevent potential virus affecting the shares - so files sit on the VM after download until I screen and move them over). How would I move data from the passthrough storage disc to the server share itself? If I map a drive - will the data flow out from the dedicated VM NIC thru the VPN and back thru the router into the unraid NIC. I'd rather not and have the data stay local. Thanks in advance.

I'd love to see details - pics or a video of a step by step would be amazing

Is the website in that link supposed to be a jump screen to buy the author a coffee?

No ideas? I was thinking maybe syncthing setup as a one way sync?

I currently have two nvme 1 TB cache drives. One is setup for all app data (except plex) and the other is plex only appdata, thumbnails etc. I want to add in a 3rd cache 4tb SSD that will act as temporary storage of files transferred from a different computer before mover runs at its scheduled time. I use the appdata backup v2.5 for the first cache drive. I'm stuck in what to use for the 2nd nvme and future SSD. Since the second nvme is plex data, thumbnails, etc I THINK I would want to backup it up entirely as the first time then incremental for changes. Is this correct or should it just be a backup of the whole drive every time? It would backup in my array in my backup share. I think I would want to schedule this either after any change or look for changes every hour and back up if true. I have to decide whether to have the 4tb SSD run mover every hour to the array or every night. If it's every hour then I guess I don't need a backup for it. If I decide nightly then I do need to back up it every ??4 hours. What app(s) should I use to set this up? Thanks in advance

I ended up going with 1. Thanks everyone

I'm wondering if I can break the raid in the synology then try to mount again. The drive is full. 10TB will take awhile via network

Anyone successful with pulling data off of a synology. I have a ds713+ formatted in a raid 1 with 2 x 10 TB drives. One drive died. I pulled the working drive and installed it in my unraid. I'm trying to mount it using unassigned disks (have it and the pro running) but it won't mount? I've read other forums and unassigned disks should be able to read ext4. Any ideas or trouble shooting advice

Sorry to refresh this - I have the AOC-SLG3-2M2 and only one drive is visible as well. Is this a known issue?

Can I take my synology drive and mount it in my server using the unassigned drive plugin and copy the data over using krusader?

Back to this topic - after some building issues I'm finally ready to setup my unraid server. Is there a preferred order for setting up the tweaks? I assume to setup the tweaks prior to installing plex: Setup these up as cache only with 100GB min free space: appdata : /mnt/cache/appdata domains : /mnt/cache/domains docker image : /mnt/cache/system/docker/docker.img Then install plex. Is there anything else in the plex setup that needs to be tweaked to set the correct location of the thumbnails, appdata etc? How important is installing the Tweaks plugin and increase the: fs.inotify.max_user_watches ? I will have 20 TB of moives/TV shows/music once I copy the data over.

I assume a random user on YT. I figure the p2000 should be enough. Any thoughts regarding my other questions?

This is a great thread. Are there any issues with running some of the above ideas concurrently? IE - Hardlinking for the storage folders in the pool and the Direct disk access for the cache? I've read a bunch of forums but its hard to find the "best method" answer. With 2 x 1TB NVMe drives is it better to dedicate one for Appdata chache and the other for separate Plex app data/metadata/etc cache? I know I would lose parity but I would use CA Appdata Backup/Restore plugin for backing up both drives. Or is it better to keep it as RAID 0 and all files on the single cache? My server will be 90% Plex with the odd UHD disc being backed up with makemkv and backing up of computers in our house. No other VMs. Last question - is there any benefit of combing Nvidia GPU and RAM transcoding? I have a p2000 that will be for transcoding and 64 GB of ECC. I will have a max of 4 video steams at any given moment. There was a comment in the ibracorp youtube video - Managed to combine this RAM technique with GPU transcoding enabled and has significantly sped things up. For anyone else wondering, to enable GPU transcoding (assuming you have an Nvidia card installed) you must first add the extra parameter "--runtime=nvidia" (without quotations) and then separated by a space from the RAM parameters/any other paramaters. You'll of course also need to have set up the appropriate drivers which you can find in Community Applications. A working config will look a little something like this "--runtime=nvidia --device=/dev/dri --mount type=tmpfs,destination=/tmp,tmpfs-size=20000000000 --no-healthcheck". This will still use your GPU to transcode but will instead offload that data to RAM instead of smashing your CPU and Cache resources. Happy homelab-ing! Edit: I should add that I have this tried and tested on linuxserver's repository but YMMV. "--device=/dev/dri" is only for CPU transcoding. Since a recent update, the container will fail to spin up if you use this with GPU transcoding, so just remove it to get "--runtime=nvidia --mount type=tmpfs,destination=/tmp,tmpfs-size=20000000000 --no-healthcheck". And of course, adjust "tmpfs-size=xxxxxxxxxxx" to be appropriate for for how many bytes of RAM you're able to allocate in your server or you might exceed your total RAM which may lead to containers/vm's etc. starting to crash.

Hmmm...did I ask too many questions for a single thread? @casperse @mgutt You guys posted a lot regarding my questions. I would greatly appreciate your opinion

Hey All, Thanks in advance for the answers/help. I've been watching/reading all that I can about the optimal Plex setup. It will be the main docker used for this server. I have 2 1TB NVMe drives for a planned pooled raid 0 cache I guess I have five questions. 1. Combining Nvidia GPU & RAM for transcoding? I watched and see below for the specific comments I pulled: Managed to combine this RAM technique with GPU transcoding enabled and has significantly sped things up. For anyone else wondering, to enable GPU transcoding (assuming you have an Nvidia card installed) you must first add the extra parameter "--runtime=nvidia" (without quotations) and then separated by a space from the RAM parameters/any other paramaters. You'll of course also need to have set up the appropriate drivers which you can find in Community Applications. A working config will look a little something like this "--runtime=nvidia --device=/dev/dri --mount type=tmpfs,destination=/tmp,tmpfs-size=20000000000 --no-healthcheck". This will still use your GPU to transcode but will instead offload that data to RAM instead of smashing your CPU and Cache resources. Happy homelab-ing! Edit: I should add that I have this tried and tested on linuxserver's repository but YMMV. "--device=/dev/dri" is only for CPU transcoding. Since a recent update, the container will fail to spin up if you use this with GPU transcoding, so just remove it to get "--runtime=nvidia --mount type=tmpfs,destination=/tmp,tmpfs-size=20000000000 --no-healthcheck". And of course, adjust "tmpfs-size=xxxxxxxxxxx" to be appropriate for for how many bytes of RAM you're able to allocate in your server or you might exceed your total RAM which may lead to containers/vm's etc. starting to crash. Is there a reason to combine Nvidia and RAM transcoding? I have 64GB of ECC and a P2000? 2. Hardlinking Another cool looking video Is this the better way to setup shares than what spaceinvader one has demonstrated in his general setup video? See question 4 below - will hardlinking conflict with this? 3. Cache & Plex- I must of read for hours regarding this. With 2 x 1TB NVMe drives is it better to dedicate one for Appdata and the other for Plex app data? I know I would lose parity but I would use CA Appdata Backup/Restore plugin for backing up both drives. Or is it better to keep it as RAID 0 and all files on the single cache? 4. Direct disk access Are there risks with this? It seems like it will really speed things up. 5. All of my UHD/DVD backups are on a synology. When I move them over (through network copy past of a mounted drive), is it better to avoid using the cache in order to prevent unnecessary writes to the NVMe drives? Ie direct writes to the storage pool? Have about 10TB of data to move over. If the answer is yes; is there a simple way to do this (ie temporarily turning off the cache for the pool)? I read some users suggesting using a larger SSD as the target drive (unmounted drive plugin) and then move over to the pool? Thanks again for all the help/suggestions

Figured it out *2M2 is for M.2 and 2E4T is for U2

I ended up getting a great buy it now deal on the X10DRi-T. I'd like to bifurcate one of the x8 ports to x4x4 for 2 x NVMe 2TB cache drives. There are two different supermicro cards that do this. Is there a difference between the AOC-SLG3-2M2 and SLG3-2E4T ??

think I'm going with the Supermicro X10DRi or X10DRi-T. I will use the AOC-SLG3-2M2 for bifurcation of an x8 to x4x4 for dual NVMe 2TB cache. Has the dual NVMe cache bug been fixed? Now which processor? Found a good deal on E5-2673 V3. Are these overkill?

Thanks! That was an awesome reply. Edit...follow up I see that supermicro MB has a physical x16 but it is electrically x8. THe quadro 2000 needs an electric x16 port. Also, I couldn't find if one of the x8 ports has bifurcation to support a card with 2 NVMe at x4 on it. Back to the drawing board

So I've done everything a bit backwards. I've gotten a bunch of parts purchased except my MB, processor, RAM and NVMe. Unraid server with plex and makemkv docker. Eventual windows 10 pro VM Case: Supermicro CSE-836BE16-R920B with BPN-SAS2-836EL1 HBA: Dell h220 flashed in IT mode (Aka LSI 9207-8i) - will connect 2 ports to the backplane and I believe an HP or intel expander is not needed 10G SFP+ card: Mellanox CX312A ConnectX-3EN Dual Port 10GB SFP+ MCX312A-XCBT GPU: quadro p2000 (transcoding) Drives: data 10TB x 10 WD elements, parity 12TB x2 WD easystore 4kUHD slim drive (For MakeMKV): LG BU40N Cache: Want dual 2TB NVMe - need MB to accomodate or card with bifurcation of x8 port So by my math I need a single PCI-e x16 (dual width on MB to accommodate size of GPU??), Three PCI-e x8 ports for the HBA, 10G SFP+ and NVMe bifurcation port I'd like to stay somewhat power-saving so I don't need a threadripper but I'm open to AMD or SP/DP Xenon. I've looked at the supermicro boards x9 and x10. IPMI would be nice as I plan on this being headless. Would love to hear suggestions or if someone has a board with processors to sell....

Does it need to be flashed to IT mode?

Which is the safest? If I accidentally download malware/randsomware and it gets past the scanning software, would either of the two suggestions protect the main server and netetwork? Is is possible to setup the connection as one way traffic where the unraid pulls content instead of the VPN download server pushing content?

I have a router (ccr) that has two SFP+ ports and multiple ethernet ports. Three of the ethernet ports are setup as direct VPN access with a different network subset and no access to the home network. I use the windows 7 computer connected to the VPN port as a download server. I'm trying to figure out if I can get unraid to look at the download server copy the files over to the unraid media share. Because I have switches with open SFP+ ports I'm going to utilize them for 10GbE speed. Move VPN connection from the ethernet port to one of the Mikrotik ccr sfp+ ports. Connect Cisco sg350 SFP+ to the other ccr SFP+ port. Connect Cisco sg500x to sg350 via sfp+. Connect the unraid storage server to open sfp+ port on sg500x. There is no layer 3 switching on the cisco switches. The mikrotik is doing the routing and switching. I need to be able to access my VPN download server from my workstation (10 Gbase-T connection to Cisco SG350) so I've been told to use Remote Desktop and have a rule on CCR router that only allows traffic from workstation to the RDP port on VPN download server for control instead of using teamviewer. I think I can figure this out myself. This would allow me to control the VPN download server but not transfer any files from the VPN download server to my workstation I think?? How to I setup a share between the unraid storage server and the VPN download server? Both are on separate subnets and there are firewall rules preventing the VPN download from accessing anything on the home network subnet. I think I have to modify the firewall to allow the unraid server access to the VPN download server via LAN without exposing it to the VPN internet traffic - need to study more Router OS to figure this out. If the firewall gets modified correctly then I should be able to use the unraid krusader docker to mount a share of the VPN data drive?? What makes more sense - mounting the share on the unraid server or on the download Server (windows 7)? I'm really not sure what to do here and don't want to invest a few thousand dollars in building a new unraid server without knowing if it is possible. Thanks for your advice

Current network https://content.invisioncic.com/r242699/monthly_2019_11/1756544085_ChannanNetworkupdated.png.98e2de3f5aada5cdeed1fb00fb3ae66d.png