Dlraven

This guide describes how I installed and configured Unraid on a QNAP TS-873A (8-disk NAS) with an attached QNAP TL-D400S (JBOD enclosure). I'd seen a few questions about this configuration, but not a definitive answer so I'm hoping this guide answers those questions. I've indicated some optional hardware changes as well. This guide is specifically for the hardware listed above and may not be applicable to all QNAP hardware. The TL-D400S is an optional component and items related to this can be ignored if not using it. Required Hardware A supported video card. The 873A does not come with built-in video. You will need to check QNAP's site for compatible video cards. Space is very limited. The PCI connector is smaller than what you are expecting and is open on the end to allow longer cards (it's really strange). I'm running an Nvidia Quadra P620. The video card must be PCI powered. A QNAP TL-D400S enclosure with expansion card and direct attached cable. A USB stick with Unraid installed. Hard drives. Make sure these do not include any previous versions of ZFS, Unraid shares, etc. If they do, format them before proceeding. Optional Hardware 2x 120mm Fans 4-Pin (I replaced the installed fans with Noctuas) 1x 60mm Fan 4-Pin 64GB ECC Memory (2x32GB DDR4 2666 PC4-SODIMM) M3 Screws (I replaced the case screws with knobs for easier removal) 2x Gen3 NVME drives with heat sinks Installing Hardware Remove the 3-screws from the back of the case and remove the top. If desired, you can replace the standard fans with upgraded ones. This is optional. Install the video card, expansion, and optional NVME drives. You may need to remove the power supply to get at the expansion slot screws. Close everything back up. Connect the cable between the 873A and D400S. Install your hard drives into both devices. On the back of the D400S, verify the fan switch is set to Auto and the Beep option is enabled. Plug in the Unraid USB stick. BIOS Changes After connecting a monitor, boot up the 873A and start pressing the DEL or ESC key as soon as the monitor receives a signal. Eventually, you'll see a standard boot up screen and put into BIOS. There's not a lot here, but you do need to check a couple items. Advanced => Hardware Monitor => Smart Fan Function => Auto Chipset => BIOS Beep Function => Enabled (Press DEL/ESC after this beep to get into the BIOS) Boot => Boot Option Priorities (Pick the Unraid stick. You'll see several USB options to boot from, so ensure you select the correct one.) Startup Procedure Turn on the D400S and wait for it to beep. It's important to do this step first and to wait for the D4OOS to completely boot up. Turn on the 873A and wait for the Unraid screen to appear. At this point, you should be in Unraid and have the option to configure your array. I'm not going to cover that here. You should be able to see the drives in the external enclosure as well. Configure Unraid Modify the Syslinux configuration boot option (Main => Flash => Syslinux Configuration). Update the Unraid OS option and reboot. kernel /bzimage append initrd=/bzroot acpi_enforce_resources=lax Install the QNAP EC plugin from Community Applications. With these changes, you should now have access to all 3 fans in the system. It does appear the two 120mm fans speed are linked. This should be all you need for a basic configuration of the 873A with an attached enclosure. Lessons Learned and some QNAP Oddness Do not install the Auto Fan Control plugin. The CPU is passively cooled and resides in a separate area with the NVME drives. The Auto Fan Control plugin is tied to drive temp which means even though the CPU is cooking, unless those NVME drives are hot, the 60mm fan will not kick in. Leave the cooling to the hardware. If someone has a way to control a fan based on either the CPU or drive temp concurrently, please let me know and I can test it. The drives in the 873A take a very long to time spin up. Like 60+ seconds. No idea why, but they do spin up eventually. Drive blinking is odd. The 873A drives lights blink when accessed, but the D400S do not. This bothers me way more than it should. Do not expect to run VMs on this machine. Due to the limited graphics support, I'm not sure the system could handle it. Dockers seem to run fine including Emby (although I wouldn't attempt multiple streams with this setup). If you install the Unassigned Devices plugin, you will see 7-drives with names like QTS_BOOT and random numbers. Don't touch these. This is how the QNAP boots if the Unraid stick is not installed. If you ever want to run the QNAP software, simply remove the Unraid stick. I don't know if you can migrate ZFS pools created in Unraid to QNAP's OS.

Already done. Screenshots are a wonderful thing.😄

This corrected the issue. Thank you very much for the help.

I rebooted, imported the pool and the resilvering started. After 8-hours or so later, the process completed with no errors. Unfortunately, I'm still unable to start the array. I'm still getting Wrong Pool State zfspool - invalid config when attempting to start the array. Forcing the pool import and then status is listed below. I am able to access the zfspool directly from the terminal and was able to back up the one folder I needed (hooray). root@Tower:~# zpool status pool: zfspool state: ONLINE scan: resilvered 2.05T in 19:04:57 with 0 errors on Wed Nov 29 16:31:48 2023 config: NAME STATE READ WRITE CKSUM zfspool ONLINE 0 0 0 raidz1-0 ONLINE 0 0 0 sdc1 ONLINE 0 0 0 sdm1 ONLINE 0 0 0 sdb1 ONLINE 0 0 0 sde1 ONLINE 0 0 0 sdh1 ONLINE 0 0 0 sdd1 ONLINE 0 0 0 sdn1 ONLINE 0 0 0 sdi1 ONLINE 0 0 0 errors: No known data errors

Here is the output of the import and status. root@Tower:~# zpool import zfspool root@Tower:~# zpool status -v pool: zfspool state: DEGRADED status: One or more devices is currently being resilvered. The pool will continue to function, possibly in a degraded state. action: Wait for the resilver to complete. scan: resilver in progress since Tue Nov 28 21:26:51 2023 687G scanned at 14.6G/s, 252M issued at 5.35M/s, 17.3T total 0B resilvered, 0.00% done, no estimated completion time config: NAME STATE READ WRITE CKSUM zfspool DEGRADED 0 0 0 raidz1-0 DEGRADED 0 0 0 sdc1 ONLINE 0 0 0 sdm1 ONLINE 0 0 0 sdb1 ONLINE 0 0 0 sde1 ONLINE 0 0 0 sdh1 ONLINE 0 0 0 sdd1 ONLINE 0 0 0 sdn1 ONLINE 0 0 0 replacing-7 DEGRADED 0 0 0 17991041654255577759 UNAVAIL 0 0 0 was /dev/sdi1/old 4345501486732808471 FAULTED 0 0 0 was /dev/sdi1 sdi1 ONLINE 0 0 0 errors: No known data errors

I had a drive fail in my zpool. The drive was still under warranty so it was sent back for replacement. I put the new drive in and Unraid attempted a resilver and failed (locked up the server and all I could do is hard reboot). Now, Unraid will start, but will not start the zfs pool. When attempting to start, all I received is a pop-up with Wrong Pool State zfspool - invalid config (zfspool is the name of my pool). Looking at syslog, it appears that zfs cannot find the replacement drive. Nov 28 21:34:27 Tower emhttpd: import 37 cache device: no device Dropping down to terminal, doing an initial zpool status check yields "no pools available". If I import the pool and then run the status, I get NAME STATE READ WRITE CKSUM zfspool DEGRADED 0 0 0 raidz1-0 DEGRADED 0 0 0 sdc1 ONLINE 0 0 0 sdm1 ONLINE 0 0 0 sdb1 ONLINE 0 0 0 sde1 ONLINE 0 0 0 sdh1 ONLINE 0 0 0 sdd1 ONLINE 0 0 0 sdn1 ONLINE 0 0 0 replacing-7 DEGRADED 0 0 66 17991041654255577759 UNAVAIL 0 0 0 was /dev/sdi1/old 4345501486732808471 FAULTED 0 0 0 was /dev/sdi1 sdi1 ONLINE 0 0 0 I cannot remove the new drive from using the drop-down on the Main tab as that yields the same error as before. Is this fixable? If I've borked the pool, is there a way to temporarily access the data? The data is backed up, but I think there is one directory that hasn't been backed up recently. I've attached my diagnostics file. Any help or directions you can point me in would be greatly appreciated. tower-diagnostics-20231128-2147.zip

So, TrueNAS can experience the same issue and their forums had a suggestion; add nomodeset to the SYSLINUX.cfg. I tried that and I'm able to get past this issue. The TrueNAS folks state that Linux "seems" to have issues with multiple video cards. I have an AMD processor with integrated GPU and an Intel A770 running on a ASRock X570 Taichi. There is no way in the BIOS to turn off the AMD GPU-- you can only set the default device and whether you want hybridgraphics. The only thing I can think of is there is a bug/change in the AGESA Combo V2 PI 1.2.0.8 microcode. I updated the BIOS a few weeks ago and have rebooted several times without issue, but that's the only recent change and not sure why this issue popped up with this restart.

Same issue. I don't suppose there is a debug mode/log file that will tell me what the boot process is attempting to check before checking it.

I've had my system up and running for several months now. I pre-emptively shut down my server due to some storms in the area. When I turned the system back on, I'm greeted by Unraid not finishing the boot process. The process is stopping at amdgpu: Topology: Add CPU node The same issue occurs when regardless of which boot option I select. I've recreated the boot drive both from My Servers backup and stock and the issue occurs. Tells me it's probably a hardware issue. Any ideas on what the boot process is hanging on?