Mattaton

I seem to be going from bad to worse. All kinds of seemingly unrelated issues. I have fully cleaned out MIAB several times. Here is what's happening... After creating the VM with the MIAB helper script, I then try to run the VM to install the OS. I get this: That led me to realize that the Macinabox_VARS-pure-efi.fd file was not being downloaded. So I downloaded that from GitHub and manually added it to the custom_ovmf folder. That let the VM start. But then I get this: As I said, I've fully removed MIAB, the VM, appdata, Docker template, etc. I've rebooted my server. Other VMs are running fine. I'm not sure what's wrong here. First time I've seen this. I've installed MIAB yesterday and created VMs, installed MacOS, I just couldn't get it to boot. I started over today and now I can't even get into the VM. 😞

I did it as @ghost82 described in his post that @KittenMittons said worked for him in the comment I linked to previously. Here's the original comment with the steps: I must be missing something in what his Step 7 means. (Again, I am a straight-up newb.) Does mount partition and change SMBIOS mean the step where we copy the EFI files to the MacOS disk? Now I'm seeing that his step 8 says NOT to copy the EFI files, but I don't know what else was done in Step 7. What does changing the SMBIOS look like? What part of SpaceInvaderOne's video did that? Sorry, I'm being dense. 😄

I swapped out the opencore.img file as described below and still have the same problem. 😕 Oh well. I'll just wait for a new version of MIAB and then see what happens.

Looks like @ghost82 has already pointed us (and others) in the right direction here: I'm going to give it a shot this evening.

I guess a third choice would be to wait for the PRs to be merged, uninstall MIAB as SI1 describes in his video, and then reinstall the new version. Then it should work if I follow his video, correct? In theory? I'm all for tinkering, I just haven't researched and gotten my head into all the details of installing a Mac VM. Obviously, it's much more complicated than installing a Windows VM. 🙂 Thanks!

Thanks @ghost82 First off, as mentioned, I'm new to Macinabox, Mac VMs, VMs, and unRAID...in order from least to most familiar. 😄 I tried doing a search of this thread for some text in what I was seeing, but I didn't come up with much. Once I saw that there was a base issue with Macinabox where it was installing the latest version (Monterey) no matter which version you chose to install, and that MIAB wasn't set up to handle Monterey yet (latest version in MIAB is Big Sur), I just assumed that was the problem. I have no idea what the whole Core Configurator and all that is doing. I'm guessing it's some intercept to fool MacOS into thinking it's booting from "authentic" hardware so it'll actually boot? Again, I followed SI1's video to the letter (twice) and it didn't work, so I'm not sure where to even start to fix it. I'll have another poke around in the threads and see if I can find anything. Thanks!

Well, at least we can say we followed the videos correctly? I guess. 😄 Misery loves company! I hope he gets it patched soon. I want to play. Haven't messed with MacOS in decades!

I'm a total newb, but I'm assuming from the latest comments that we're out of luck on using Macinabox right now? Mine installs Monterey no matter what, as others have mentioned. After going through the entire install procedure twice per SI1's video, I get to loading up the VM the final time after OS install and VM settings tweaks, I get the OpenCore Boot Menu where I have to select the boot disk, then I get the attached text on the screen and nothing else happens. 😕 Hopefully he can get this tweaked soon!

Makes sense. I'll give this some thought! Thanks!

I'm not hosting the VM on unRAID. The VMs run on a Windows Server 2016 machine. We are using the NAS as a backup location. I need to be able to backup 2TB each night. Copying that data straight to the array is slow, so I want to install a large, redundant cache pool. I don't know if that's as easy as buying two 4tb NVMe drives to replace the 1TB I have now. Might be the best/easiest way, but it's expensive! 😄

Currently I have a cache pool consisting of dual 1TB Samsung 980 Pro NVMe SSDs. I built this NAS because we're overhauling our backup system and needed a NAS for that. I opted for dual SSDs for the safety of not losing a disk before Mover gets the data to the array. Anyway... turns out the 1TB cache is too small. Our backup of one VM hosted on a Windows Server 2016 server is 1.9TB and then we have two smaller VMs on the same Windows Server, plus the HyperVisor host. So, too much data for the 1TB cache to be useful. I'm currently writing directly to the array, which is taking about 13 hours. Obviously, I'd like to bump up the size of the cache, but even 2TB isn't enough. So now things start getting PRICEY. My question: Do any of you have caches larger than 2TB? And if so, what's your strategy? I have three M.2 ports on my mobo (ASUS Prime Z590-A), but one of those will continue to be a system cache drive. So, I have two ports to play with. I should also have room to add another PCIe card in addition to my USB 3.1 card (passed-thru to Windows VM) and HBA LSI card. Thanks!

Hmmm...seems some do and some don't. My system log matches, but my VM Manager (libvirt) log doesn't.

We have an outside IT firm setting up Veeam on a local Windows Server 2016 machine to backup to out unRAID NAS. They are getting errors and failures writing to the unRAID share. Can any of you gurus shed some light on the warnings in the system log at 11:22-ish timestamp? Some samba issue? Seeing things like: Dec 8 11:22:50 AP-Archives smbd[27615]: smbXsrv_open_close(0x69ee87f0): failed to delete global key '69EE87F0': NT_STATUS_NOT_FOUND Dec 8 11:22:50 AP-Archives smbd[27615]: smbXsrv_open_destructor: smbXsrv_open_close() failed - NT_STATUS_NOT_FOUND Dec 8 11:22:50 AP-Archives smbd[27615]: smbXsrv_tcon_disconnect(0x4ddc1133, 'BU_APServers'): failed to delete global key '4DDC1133': NT_STATUS_NOT_FOUND Some further info: We've been having trouble getting the backup to write to the original share I had set up for it. This was all due to the fact that the backup consists of one large ~800GB file. The cache is 1TB and the original min free space on the first share was set to 50GB. So, the 800GB would start to write to cache and then run out of space. I created a new share with a 900GB min free space hoping that would solve the problem. And now I have it set to not use cache at all. Not sure if these errors still have something to do with all that, but sharing the info just in case. Thanks!!! ap-archives-diagnostics-20211208-1141.zip

From reading, I guess this is normal behavior as Linux resets the BIOS time to UTC, but has anyone figured out how to make the times in log files match the unRAID system time? I'm assuming the log files are written using the BIOS time (UTC), whereas the unRAID time is set to my timezone. It's a minor annoyance, I suppose, but it can be confusing trying to analyze log files when the times are all wrong.

Agreed. It's just tough to perform such upgrades when the server is so often in use. All these employees start complaining really fast when a server is off. 😄

This was coming from an outside IT guy we have installing/configuring the backup software. When I map the share as a drive on my workstation, it does report correctly. I'll have him check again to see if the size reports correctly.

There is a BIOS update from July. Current BIOS is from April. I'll try updating that some time. When I do, should I just remove the "pci=noaer", reboot, and then try to start the VM?

Well, that's not what I wanted to hear! 😄 Not sure at all how to tackle that issue!

The folder is already on the disk. There are some files from the share still on the cache, but the folder within the share I have set as the backup destination is completely on the disk.

Thanks, @doesntaffect! I'm going to see how the U25 behaves before I replace it. But yeah, if you can send me the details on the U26, that'd be great! I'm curious which chipset it uses versus the U25. I wouldn't think they'd completely swap chipsets, but maybe so!

I have a share set to "yes" on cache usage. We have backup software running on a Windows server set to use the NAS share as a backup destination. It reports the destination as only having 980GB free, which is the size of the cache instead of the 3+TB available on the disk assigned to the share. Is there a way, other than not using cache, to have it report the "real" available space? Thanks!

Yes, I believe that's the one SpaceInvader One used, but I think I saw that the Asmedia would work, but the chipset had a different number. Like 1142 instead of 2142. Something like that. I took a chance that I was just getting the latest version of the same controller and it'd work just as well. Not quite true. 🙂 Thanks again for all your help. VERY much appreciated!!!

Interesting. I'm using an Inateck Red Comets U25 (has three type A and two type C ports). I guess they used a different controller on the U26. https://www.amazon.com/gp/product/B096ZK6C37

Dang. I was hoping to get one that worked well. I was trying to find one on Amazon that SpaceInvader One was using, but I couldn't fins the exact model he was using. It's also difficult to tell in the description and comments which controller is on the card. If it gives me any trouble I will try to find a better one. I need one that is USB 3.1 with type C, so I couldn't just choose any card I found.

Thank you so much for your patience in showing that to me. And THAT WORKED! I have no idea (yet) what the pci=noaer did, but it passed the controller perfectly. Thanks so much!!!! PS: I still see new virNetSocketReadWire errors, but if everything works, I'm just gonna ignore them. 🙂