RAP2

So my parity process completed - and everything is fine. However, this process of creating a Windows VM has caused me to second guess using UNRAID. I downloaded Virtual Box and created a Ubuntu and Windows VM pretty easily with that. In the space of just over a week I had the UNRAID system become "unavailable " twice; due to two separate and IMHO, normal operation: start a VM and "Compute" share spaces. The former required me to restart the machine - and then it needed to spend 10 hours rebuilding the parity. This, with only a test migration representing 1 of 8 drives... Some might argue "beginner problems" - that's fine - but I thought UNRAID was built for home users; not data center engineers. I don't *want* to be an UNRAID expert; I don't have the time. Very little of my UNRAID experience - including getting the USB flash drive to work, was simple and straightforward. And while searching for solutions on this forum, I was not alone. Very early in this thread, I mentioned that I noticed some performance issues with user shares. This was easily observed when Windows Media Player was rebuilding the database for my 800 GB music library; it was much slower then the original NTFS/Win10 implementation. I redid my Windows 10 just the week before, so it was fresh in my mind. I started doing research on that problem and found this: Unraid User-Share vs. Disk-Share SMB Performance During that research I also discovered SNAPRAID, which advertises similar features to UNRAID - example: parity drive(s) with heterogenous drive sizes/models. To be sure there are pros and cons - but I like the fact that it has integrity and silent error management. Its not real-time but I can create scripts to schedule snap-shots and data scrubbing. Best yet, I can use NTFS so there is no data migration required. Performance is what I'm already used to. I want to thank all the folks that participated in my UNRAID education; in particular, Squid, itimpi, JonathanM and Frank1940. Its folks like you that make the on-line tech support systems work. I appreciate you. 🙂 Best...

I'm curious what caused this. Here ya go... mediaark-diagnostics-20211023-1528.zip

Linux Guide shell says nothing about restarting or shutting down a machine, so I did a hard reset. System booted but notifications popped this up: Parity Check Tuning: 23-10-2021 11:05 [MEDIAARK] Automatic unRaid Correcting Parity Check will be started Unclean shutdown detected Its now a 10+ hour process. 😞 Seems like trying to get a VM going on UNRAID can have some potential serious issues getting started.

OK - it might be running, but its not working. All my user and disk shares are dead. Is there a way from the local machine that I can enter into the command line to stop the array and restart the system safely?

I just tried to restart the VM and it has taken UNRAID offline. (i.e.: the Edge webGUI timed-out with: "Hmmm… can't reach this page") Now what? I've not had that problem in the past. The VM would start - I just could never connect to it via RDP or VNC. Thats the second time I've lost WebGUI connection to UNRAID - the last time for a couple of hours because I pressed "Compute" space in shares. Is it possible, something in the VM config has gotten corrupt with all my various attempts? This is my second version only created yesterday. I removed the first attempt. Ideas? I've now tried to refresh the page 3-4 times over 10 minutes. I walked over to my server room and just looked at the local modnitor; seems to be still alive: the linux text and UNRAID login at the bottom with a flashing cursor.

Thanks itimpi. Isn't that path when you want a vDisk? SpaceInvaderOne's directions was pretty clear on that. He mentioned that an NVMe is not a drive but a PCIe device; and to leave that vDisk location filed set to "none". Having said that, his instructions are not working! 🙂 I guess I just want to be clear on my particular circumstances before I do something that might destabilize the system. I have a Windows 10 Pro installation on the primary partitions of a 3 partition NTFS NVMe drive, installed on the MB. That drive boots into Windows 10 when the URAID USB is removed. I'd like that to be the case - but still be able to run that Windows install as a VM within UNRAID. SpaceInvaderOne's video was specific to that application. Only problem is that he did that all in an older version - for example - the whole PCI binding thing is different now. Happy to try it your way - but before I do, I'm wondering if its a NIC thing. When UNRAID boots my primary MB NIC is sued by UNRAID. Does this same NIC get virtualized and so the same IP address is used by Windows? I bought a separate 2-port Intel Server PCIe NIC specifically for the 2 VM's I want to run (Windows 10 and Ubuntu). Its just not clear how to assign a card to a VM. Right now I have the following. VM: Running a VM - along with the UNRAID parity system - are the two main reasons I'm moving platforms. But if I cannot run a VM, easily under UNRAID, I can try soemthing else. It's one reasons I just wanted to get one Data drive migrated and tested before I make decisions. But I do want this to work!

Well I decided to wait it out - in case the copies were still being made in the background. Glad I did. Two hours later, I checked on the machine and I was able to remotely connect via the WebGUI. The copy was not done but over night it completed. I guess I won't be pressing the "compute" link again... Someone should look into what's going on with that; it should not take the system off line for any amount of time; IMHO.

I should also note that my user and Disk Shares via Windows machine are broken - times-out with "Windows cannot access \\UNRAID"

Quite unfortunate, since I was in the middle of coping 8 TB's of data in a migration. I have no experience with an UNRAID (apparent) crash. My EDGE web browser shows: "Hmmm… can't reach this page" But is that the UNRAID web server - or UNRAID itself? A power cycle could cause more problems if UNRAID is still copying files; and it should be for the next 6-8 hours, based on my last estimates. I also went to the machine itself and the screen is blank - normally it has a bunch of Linux stuff and the login at the bottom. Some advise would be appreciated it. IMHO, Looking at the cat should not kill it...

Here are my current settings... VM:

I ended up removing that VM and started from scratch. Here are my current settings: VM:

New VM Settings:

I've deleted the VM and started again... I still have br0 missing. Is there a problem using virbr0 instead? In any case, same problems: cannot RDP - or VNC into the system. When VNC loads, it shows the UEFI Interactive Shell window... only thing I could do there is type exit. Then it went to an old school BIOS setting window. Reset or continue does the same thing; circle back to that UEFI shell... its in a loop. This should be easier IMHO.

I've recreated the VM - taking into account the GUI approach to creating the stubs, after reading this post: Guide: bind devices to vfio-pci for easy passthrough to VMs - VMs - Unraid The VM starts - and although I set up Windows 10 to handle RDP, it will not connect. I tried changing the VM to use VNC and when it loads, it shows the UEFI Interactive Shell window... only thing I could do there is type exit. Then went to an old school BIOS setting window. Reset or continue does the same thing; circle back to that UEFI shell... its in a loop. What does THAT indicate?

Not sure where that file is. I used Krusader to look at the tmp directory: It does not exist. I'm confused. I did not check to use the led controller in the VM. Why would it attempt to parse it?