methanoid

Still barfing.. went without ANY VM or Dockers.. barfs in around 10m of trying to do parity check. Can't grab logs cos the machine is locked completely. Disabling parity check and running VM again. See if it locks at all or if its the creation of parity that is the problem. Does LT do any support or do they rely completely on the Forum doing the job for them?

Still barfing... removed GTX970 and fitted crappy HD5450 (should eliminate any question of PSU draw!). I wonder.. I have been watching youtube when its gone all dead... wonder if its to do with audio issues? No ACS Override on?

8 hdd. 550w silver stone gold psu (SST-55G). Never any issues before with "other" products. It's just barfed again. Will have to try tests. EDIT: Mem tests fine.. have rebooted and disabled all Dockers. Just unRAID and Win10 VM... If that fails will just do with unRAID on own...

Damn... OK.. I'm stumped then!

I may have solved my own problem... I had one disk at 98% full... I didnt see that unRAID has ALL hard disks defaulting to Warning disk utilization level (%): at 70% Critical disk utilization level (%): at 90% So presumably at 90% of parity completion it then barfs? Why would a default setting mean you cannot use the capacity (or even close) of the drive. For a 5Tb drive that means I am expected to leave 500Gb empty? And would I want a warning at 70% of capacity (when 1.5Tb is free)? I have changed the values on each HDD and set the Parity drive "critical" level at 1% higher than the Data drives. Fingers crossed that is the issue. Parity sync again... See what happens in 14 hours or so

I have a Win10VM (setup as per Wiki so no power management issues) on top of unRAID. I come down this morning and screen is blank (like time out from VM) and mouse movement makes no difference. It is "dead" But the VM hasnt passed back the control to the Console so no KB & Mouse there. Try login to WebUI on phone (no fun on 4in screen) and no response. No choice but to reboot. So I still have so far failed to create parity. Twice now... And could not capture any logs either. Wondering why parity creation seems to lock the machine? Machine has been completely stable in "other products"

So what happened? I've been trawling forum for info on how to set up CPU pins.... Similar to you... 4 core, 4 HT... Some Dockers inc Plex ... Win10VM for games - Wondering what I can assign to what or if I need to mess with them? Could I run two GPU-assisted VMs?

OK...if I have unRAID console on one monitor and boot a VM on another.... and want to switch to a different VM. I cannot do so from the web interface as I'd need to shut down my VM first and then I lose the web interface... So, can I start/stop VMs easily from command line? Ideally without LOTs of command-line-fu, maybe a batch file to make it less typing? Something like Win10 stop, Linux start, Linux stop, Win10 start Can't see anyone else asking this but if I can it will make life a lot easier (and cheaper!) EDIT: Seems I can with "virsh shutdown Win10" and "virsh start Linux" SO.... I guess now (for minimal typing!) the question changes to IS there a way to alias "shut" to "virsh shutdown" and "start" to "virsh start" ? Sorry... Not really used to Linux !!

Did you ever get an answer? I have just started adding Dockers to my setup and was puzzled to see no port numbers etc in config for the web interface. Is that not configurable or is this Docker broken? Port Mapping is only available when Network type is Bridge. Also try the Advanced Settings slider in upper right. Thanks, a little embarrassed as I thought it was missing on 2 dockers and it was in Adv Settings... I'#ll learn!

Thanks.. didnt spot it until I switched Adv Settings on Oooops, sorry, noob here!

Yeah I am not "worried", just a little OCD about it I think A bit like when I keep getting warnings about my Docker image being 70% full... why tell me then? Tell me when its 90%!! I'll end up expanding the image just to stop the warnings!!

@Sparkly, does your "phusion/baseimage:0.9.16" differ from the "linuxserver.io" base image by much? I know you are part of the LS.io team but you also have a separate repo so maybe they do?

Q: how different is the phusion/baseimage:0.9.17 to linuxserver's base? Just looking for dockers which minimize the extra pull where possible. Q: Mylar docker.. no mapping for port number? An oversight? Does that mean I have to use 8090 default only?

Did you ever get an answer? I have just started adding Dockers to my setup and was puzzled to see no port numbers etc in config for the web interface. Is that not configurable or is this Docker broken?

I wonder if it is possible to script the VM startup script to shutdown (cleanly) and other VM using the same GPU. Otherwise with one machine how can you close one VM and open another?

I have found that it is somehow related to the GPU used. I have a Gigabyte DS3P with two x16 slots (The second only x4 mode). I have two GPUs, one Gainward GTX 760 and one Sapphire HD6450. I can not get the Nvidia card to work if it's in the first slot, but in the second it works if there is a card in the first slot. If there is no card in the first slot, it doesn't work. The AMD card works in both slots no matter what. But when 6450 is primary and 760 in x4 you cannot pass the 6450 to a VM (and maybe also the 760) Your GA-970A-DS3P doesnt have IGP. If you cannot pass the 6450 that effectively confirms that IF I get an X99 mobo I would need a crappy card for console (in primary slot) plus any GPUs I want to pass.... Sorry if I am labouring the point a little!!

Is it however possible to start a VM from the console? Yes I know you can autostart a VM... but.. Example Scenario unRAID running on IGP for console, GPU passed through to Windows VM. Want to switch from Windows VM to Linux VM (or Mac!) Does the starting of a VM from the web interface using the GPU currently in use automatically cause the Windows VM to cleanly shut down and then the Linux VM to start? If it does, then problem is solved. If it does NOT then you get dumped to nothing on the GPU screen and have a console only to start your Linux VM. I ask as I hope to run VMs for Windows, Mac, AROS, Android etc and would like to be able to easily switch or even to know how to switch!! Thanks EDIT: I found you can start a VM from SSH with “virsh start VMname-script??” but don't know the names. Hoping I still don't have to as someone might know if starting a VM with same GPU might power down the other using it?

What base image are these Dockers based on? Is there an easy way to tell for other Dockers which image they are based on. Ideally more efficient to get Dockers all based on same base image?

Interesting... a little bit of a lottery then. Only safe way to run 3 diff GPUs for VMs is to either have onboard GPU or a 4th GPU (in primary slot, so a mega mobo needed!) ?? What about adding a desktop and browser to the unRAID console so at least that GPU could do something useful rather than sit around displaying a console window? I'm getting CLOSE to what I am looking for... hopefully without spending the $0000s that Linux spent on his multi GPU build...

So AMD in primary spot and nvidia in others? That is consistently working?

Cobra You have two good GPUs. Can both be passed through to (any) VMs at the same time? Or do you pass through one to Windows and use GTX970 to display the unRAID console??? My post here is also similar question if you have an answer - https://lime-technology.com/forum/index.php?topic=41665.15

I suspect some unhelpful editting has been done here as pci-stub has NOT been mentioned above (or has been removed). I was messing with KVM and pci-stub in Ubuntu before deciding I might use unRAID instead. I am looking at X99 mobos (no onboard graphics) and want to pass up to 3 GPUs. Therefore need to know if this problem got resolved at all. I need to be able to pass through ALL the GPUs present. I don't need Unraid hogging a GPU to display some text when I have a web interface I can access from any VM. Is there a solution or does unRAID really need a dedicated GPU to be wasted in that way???

Sorry, I was talking more about how I would LIKE the cache to behave and the mover script.... but yes, good advice for OP (but somewhat passed the sell by date!)

This was the thread I came to when looking for info about how the cache works (or maybe SHOULD work). What I would like to see is that the cache is allowed to fill to a certain % before being bypassed (effectively disabling cache) AND the mover script auto-run...before the cache is re-enabled..

I cannot understand why this feature is not in UnRAID as standard by now....