bonkers

Several things were done, one of them probably includes the solution as the server has been reliable for a while now. -give the 2 VM's 5 threads of the CPU instead of 6 (thus leaving 2 threads dedicated to unraid) -put the VM's clean to W8.1 (after continuous problems with W10, spontaneous update to anniversary being the last drop) -Fresh Unraid 6.2.1 USB drive in USB2.0 port, not 3.0. No further changes were made, I searched for any USB settings in BIOS but couldn't find anything regarding power management.

I deleted all files in the SSH folder & rebooted. I can SSH again but had the same problem again. I can confirm I have internet connection, IP is static, time is correct. I'll order a cable to connect via USB 2.0 internally. The mobo usb headers are covered with a GPU and the case is modded to not be opened easily as this system is built to be on expo's, which is why I'd rather avoid putting the Unraid USB it in the case. Replacing the Unraid files on the USB drive has consistently fixed the problem. Anything else I can try if that would fail? I have only short time-windows where I can work on the system. What I have tried now, to keep things clear: -Different versions of UNRAID (6.1.9, 6.2.0 (+ all RC's), 6.2.1), basic licence -When 6.2.0 stable came I did a complete reinstall of Unraid on a different USB drive (sandisk ultra fit 16GB USB3.0 + format the drives & reinstall VM's I'm assuming the files on the usb drive get damaged somehow? How would this be related to USB3.0?

It has been running normally now for last startups, I can't afford to try to reproduce the problem, perhaps during the weekend. I lost the ability to SSH though, can i check internetconnection via the webgui? SSH was working in the past via putty, anything i can do to re-enable it? I'm getting an error: network error : connection is refused.

I only have usb3.0 on the exterior of the case and it's a huge pita to reach the headers on the motherboard.

I have purchased a licence several weeks ago and server has internet connection via static IP. It has access to the internet as far as I can tell.

I keep having the problem that when I boot my machine, I can't start my array. I had this in 6.1.9, 6.2.0 and now 6.2.1. Rebooting doesn't fix it, sometimes shutting down and starting again resolves the problem, if not i have to copy unraid to my usb again to make it work. When I try to register again online in the webGUI it says ERROR(1) without additional info. I have tried changing USB drives (now using sandisk ultra fit 16GB usb3.0) and complete reinstall of unraid including VM's. I'm always careful to shut down properly & time is correct.

Software development steps (or at least how i call it): 1. After architecture planning etc and first very basic functionality is being made the first builds are pre-alpha. 2. When most of the functionality is done or at least the basic foundation is made for all wanted functionality it goes into alpha phase. 3. Beta phase; practically all wanted functions are present but stability is unsure or known to be wonky. At this point some devs choose to release to public to get an overview of what needs to be done still to make it release-worthy. 4. Releasing RC's or release candidates, functionality is at 100% and software is often made public at this point. If no terrible problems arise the candidate could be actually released with only minor adjustments. 5. RTM or release to manufacturing at this point the software is considered done or at least good enough for release to the general public. Some dev teams would skip releasing RC's after beta, depending on how they label stuff, how critical they are about their own work and how critical it is that a released version is as flawless as possible. Software like unraid will power machines that have to run reliably 24/7 so they make the actual released for public versions pretty rock-solid. Releasing something like fallout 3 is a whole different story where few shits are given about stability upon release and the team is under pressure by the investors to release what they have so they can start filling the financial pit they made during production.

To prevent issues on my machine with 2 VM's that have to run at max performance when they are running i disabled C-states and the CPU is locked on a decent overclock. I can check on cpu frequency with 'lscpu' command via SSH. From overclock.net: C states are processor idle states. The more C states you have enabled, the more aggressively your CPU can throttle down its power usage when it is idle. When it detects computer activity (a hardware interrupt), it will throttle back up to C0 (full power state). If you disable C states, your CPU will be in C0 state all the time. A Google search for C states will have more technical explanations. This website is worth a read if you have some time and are willing to do some reading: http://www.hardwaresecrets.com/everything-you-need-to-know-about-the-cpu-c-states-power-saving-modes/ Your ideal settings will depend on what kind of VM's you have, and if your system is stable when C-states are enabled. I have found that a heavily overclocked CPU dislikes enabled C-states but your mileage may vary. Best way to test performance in your particular usage scenario is to actually test it on your system

Problem persists

Again the same issue after running stable for 5 days.

It was the syslog i found on the stick, i couldn't manage to boot even in safe mode so didn't find a way to run diagnostics. I was thinking there might be some relevant info in there. I'm up and running again by putting 6.2.0 RC3 again on the thumb drive, array is intact as well as my VM's, no idea what went wrong. Edit: System is behind an uninterrupted power supply (APC smt1500i) so the power going in should be fairly clean.

Can't boot unraid anymore after a clean shutdown. memtest gives 2 clean passes, attached is a syslog.txt. Usb stick containing unraid is fat32 and no problems found with checkdisk. Up next is trying to go back to 4.1.9 I guess? syslog.txt

That last bit is what did it! Starting a VM with that card attached then stopping said VM made it so the input/output error no longer happened! THANK YOU SO MUCH OH MY GOD Glad I could help I think I got my first grey hair spot trying to figure out a solution.

Suddenly got it to work again after half a dozen hours of losing my mind. Here some steps I did that might help others in the future: What didn't seem to make a difference: -I went back to 6.1.9 stable but I don't think that made a difference (the only problem was obtaining the bios) -I tried several different cards from the same model & brand, as well as different slots, didn't seem to make a difference -I looked into whether there would have been some permissions problem, did not seem to be a problem -My hardware checked out fine as well, array didn't have issues as well -Tried rebooting, using fresh array, using clean unraid installs, new VMs etc -Tried extracting bios with GPU Z (gives a significantly larger bios and does not work for this purpose) and the actual bios i modded with maxwell bios tweaker and flashed onto the cards. In the end I kept horsing around with the above without obtaining a bios, still not sure what exactly did it but the following were my last steps until I got lucky: -Updated the NVidia drivers on the w10 install on the VM that will use the GPU on the first slot, using a slot that is not the first -Starting and shutting down the VM without rebooting (so that "Kernel driver in use: vfio-pci" shows up when you do "lspci -v", followed by unbinding the GPU as described above) Those last steps in combination seemed to do the trick, not sure if that actually made a difference but I'm up and running again at last and never losing my rom backup again!

Cant get this to work anymore. Using a dumped bios from another GPU with the same bios flashed onto it gives nothing but a black screen when booting the VM. Can't dump the bios anymore as well, this worked on an older version of unraid but not the latest RC3: root@UNVR:/sys/bus/pci/devices/0000:02:00.0# cat rom > /boot/vbiosdump.rom cat: rom: Input/output error Anyone who manages to extract a bios on RC3?

I'm trying to purchase a license but either the website is running on a potato or there's something preventing me from completing the purchase. I already went past giving my paypal credentials (took forever for the cart page to load properly) but it gets stuck after filling in the form with my information. EDIT: finally worked, took about 10 minutes for the page to load so this thread may be removed.

A little late perhaps but you could try getting the correct bios for your cards with GPU-z or here as earlier stated: http://www.techpowerup.com/vgabios/ What GPU are you using? I'm going to give this a go next week with 2 980ti's and i7-6800k (no onboard graphics), no room for an extra gpu for unraid. EDIT: worked as a charm, I had the same error at first, forgot to unbind the gpu first: echo "0000:02:00.0" > /sys/bus/pci/drivers/vfio-pci/unbind

It's going in a µ-ATX build with dual GPU's so only room for one extra pci-e card. I suppose I can hook up one device to the motherboard and the other to the card? The reason I'm going for unraid with 2 virtual machines on a single PC is that it seems impossible to run 2 vives since they share the same ID/vendor (and I was told adding a pci-e USB card to give completely to another virtual machine can circumvent that problem). I'll get the card you advised, thanks! (and congrats on your 999th-1000th post )

Any recommendations for a brand or is it all the same?

I am going to need 1 or 2 systems for showcasing VR tech, I have 2 HTC vives to connect. I tried both on a single machine without virtualization but it won't work due to USB issues. (Asus Rampage V extreme motherboard) For cost and mobility purposes I'd prefer using a single system with unraid, each system having a single 980ti for now (might upgrade to 1080's later). Will it be enough to get a PCI-e 4 port USB 3.0 card to connect the vive to so I can switch that entire controller to the secondary virtual system?