2025-02-16 - Active Directory 0.5 - setting up Lenovo ThinkCentre m910s with Proxmox to run Active Directory Remotely

    I never seem to get that far into Active directory before something happens and I need to start over. I got too busy with other priorities and by the time I came back to it, I had decided that I needed to buy my own server for a home lab. For the time being I'm using a Lenovo ThinkCentre M910S with 16GB of RAM. 

    I added the Lenovo to my computer stack in my room under my file server, I got it all plugged in and I'm hoping to get it set up so that I don't have to keep worrying about monitors, mice or keyboards. I want to run this thing completely remotely from my laptop. The last time I worked on Active Directory I was doing it from my GPD pocket 3 laptop on top of Windows 11, Virtual Box, and Windows Server 2025. The laptop fan was running full blast the entire time which made me think, why not just set this up on a server at my house running Proxmox? And a friend of mine who was already in IT suggested I set up a home lab in another conversation. 

    So here we are, I plugged in a thumb drive with a Proxmox installer on it, I've got the system turned on, I spammed the F12 key while booting, it went to the boot loader, I selected the USB drive, and a bright, white screen with the options to install Proxmox with the GUI, or just with the Terminal, or with the Terminal and Serial Console, or advanced options appeared. I selected Graphical since for my purposes, I do want the GUI and am used to using it in class at school. 

    The difference between Terminal and Terminal with Serial Console is basically:

Graphical = Full installer with GUI.

Terminal UI = Text-based installer but still user-friendly.

Console Version = Pure command-line install, useful for low-resource systems or remote serial console setups.

    

    I selected the GUI version and after some loading it stopped with a box saying "No support for hardware accelerated KVM virtualization detected. Check BIOS settings for intel VT / AMD-V / SVM. I had to reboot and mash F1 to enter BIOS Advanced tab > CPU Setup > enter Intel (R) Virtualization Technology and change it to enabled > ESC. 

    And because I want this system to reboot when power is detected and I have it plugged into a UPS system without ethernet so at least if there's a brownout, everything will shut off completely and not cause a bunch of errors, and when power is detected, it'll simply reboot and start running again. 

    So after pressing ESC, I went Power tab > After Power Loss and pressed enter > selected Power On. There was also an option called Automatic Power On, and under this was different power options such as wake on LAN, or from serial port ring, or various alarms. I left wake on LAN at automatic. 

    I thought, gee what else should I take care of since I'm in here? research turned up that I should turn off Secure Boot because Proxmox doesn't always play well with it, "uses custom kernel modules and drivers that may not be signed with the Secure Boot certificate. Secure Boot is designed to ensure that only signed, trusted bootloaders and operating systems are loaded, which can block Proxmox from booting properly." Apparently this is common for many Linux based distributions that  don't use signed Kernels. 

    Research also suggested turning on Turbo Mode under the CPU tab in CPU setup might be useful for particularly intense tasks but for most things in Proxmox, it is unnecessary. I was going to enable it just in case but it was already enabled. 

    I saved and exited setup, I mashed F12 again and selected GUI version of Proxmox. It started loading again. I agreed to the end user agreement. It already had the c drive selected so I hit next. The next page wants you to type what country your in, but when you type U for United States, a dropdown appear with all appropriate selections and then select your time zone. 

    Another page for management network configuration came up and it had an IP address section that made me think I should reserve an IP address on my router which then caused me to try logging into my router which I could not do. And so I had to start shooting that, I logged into my router on my desktop but I couldn't do so on my laptop which is where I do all of my productivity. It's turned into a whole rabbit hole which snowballed into about fifteen other problems such as logging out of my Google Chrome account and not being able to log back in and all I wanted to do was make sure I could log into my router when I needed to. I am just lucky now that I am able to use my desktop. 

    I went to set the manual DHCP assignment in the router and found the device in the listing, but found two devices that share the same MSFT 5 0 client name. The router automatically assigned them as static, not manually reserved. My understanding is that if you set a static assignment, that is on the end-device itself. So the router must be seeing that the end-device has it set to static and is noting that it is static and so keeping that IP for that device. I tried to set a DHCP reservation for the MAC address of the machine I'm installing Proxmox on and every time I hit apply, the reservation just disappeared and the only one still there was my file server. I tried it twice and the router refused to reserve it. I have concluded that it must not be reserving it because it already sees it as a static IP and so it will keep it. But this makes me uncomfortable that if for whatever reason this server goes down and after the lease expires, it comes back on expecting the same address, its no longer available or worse, it takes another address and I try to log in remotely using the IP for the Proxmox server and I won't be able to log in because the address I'm using to log into Proxmox is no longer associated with that end-device. Plunging forward. 

    But the second assignment with the same name has a different IP address, a different MAC address, both are connected by ethernet, both are set to static, and both use a Windows icon even though one of them is clearly the machine I am installing Proxmox on. The machine does has Windows 10 installed so maybe the router is seeing it as a Windows device till Proxmox is finished installing. And I asked ChatGPT if perhaps there's a second assignment because I started installing Proxmox before going back and turning on Virtualization in BIOS and it said its very likely. It also said that the static listing for the other device with the same client name but different IP and MAC addresses will probably clear out on its own after the routers lease time expires and the device doesn't come back on the network. I will have to test that one of these days. That didn't take long, the listing for the device I was trying to reserve disappeared and only the old one is still showing. Will have to address this later. 

    On the Management Network Configuration page I set the hostname to proxmox.ad.local because I wanted this accessible by both P address as well as by hostname. 

    And now the listing with the proper IP address has reappeared in my routers client listing page. 

    I took a picture of all the configurations for future reference and pressed install. 

    It rebooted automatically when finished after just a few minutes. After rebooting, it then went to the installation page automatically again. But I unplugged the thumb drive and rebooted and it booted up again, into Proxmox and showed a terminal instead of a GUI. It asked for the username and password and so I entered root, which is the default username and the password I gave it at set up and then looked up if I will have to locally enter the password every time this server boots because this would make remote work on any occasion the server has to be rebooted a serious problem. But my research turned up that I can SSH into the device and enter the password, since this after all is the terminal of the server that's asking for it. 

    This led me to another problem--I thought. I have used Proxmox several times in class but never in a terminal. I looked up why I was getting a terminal instead of the Proxmox with GUI that I installed and apparently this is just what happens when you use a monitor directly on the Proxmox server. Once you log into the web interface from another device, you see a regular GUI for Proxmox, which I got. 

    Proxmox is set up!

    I bet you thought this was over. No. I have an ISO of WinServ 2025 on my file server in Ubuntu server. And apparently this is a possibility, to link the file server to Proxmox. So lets roll up the sleeves and get started! 

    I went about creating an SMB share on the Proxmox server. SMB is already set up on the file server, so I had to SSH into the Proxmox server, and I entered:

    mkdir /mnt/fileserver

and I create a file called:

    nano /etc/cifs-creds

I entered:

    username=<fileuser>

    password=<filesecurepassword>

Then I entered:

    mount -t cifs //<ip address o server>/sharedfolder /mnt/fileserver -o credentials=/etc/cifs-              creds,uid=1000,gid=1000

This errored out. It said 

    mount error(2): No such file or directory

    Refer to the mount.cifs(8) manual page (e.g. man mount.cifs) and kernel log messages (dmesg)

The error indicates that either the shared folder path or the directory where I'm trying to mount it doesn't exist or is wrong. I entered:

    ls /mnt

This returned what seemed to me might be a generic name like if ti didn't know what the actual name of the file server was, so I decided to make sure this was the actual name it is supposed to have, if it mattered. So I entered:

    cat /etc/samba/smb.conf

A few days later: trying to set up SMB share between file server and Proxmox, but in trying to SSH into the device and I am now doing so remotely through my VPN, it doesn't want to SHS. So I am checking to see if SSH is enabled in Proxmox or if that is even a thing. I have access to the web interface so I am in there and took a few minutes but I was looking for the shell in the Proxmox web interface, which is located in one of the two left-side panes, the one with the tree architecture and the one to the right of that which has the specific options for each part of that tree. After realizing that I need to click on the nose in this tree, and that the node on my machine is called simply Proxmox, the first thing under Datacenter in the tree, Shell then appeared in the other pane and I clicked it. 

Now that I'm there, I am supposed to enter df -h in order to verify

A few days later, okay I just started by checking again to see if SSH worked remotely over my VPN since I am out and working on this, and this time SSH worked and I was able to get into Proxmox. I then went about trying to set up the SMB share between my file server and Proxmox so I can retrieve the Windows Server iso I have there. I entered the following  to confirm the necessary packages for SMB was installed. 

          dpkg -l | grep samba

It returned:

ii libldb2:amd64 2:2.6.2+samba4.17.12+dfsg-0+deb12u1 amd64 LDAP-like embedded database - shared library ii samba-common 2:4.17.12+dfsg-0+deb12u1 all common files used by both the Samba server and client ii samba-libs:amd64 2:4.17.12+dfsg-0+deb12u1 amd64 Samba core libraries

Looking like SMB is installed, I went to make the mount point for the file server so I can share files between the file server and Proxmox server. 

           sudo mkdir /mnt/fileserver

But it said that  

          mkdir: cannot create directory ‘/mnt/fileserver’: File exists

Basically it's already there and another command: 

          ls -l /mnt

That showed 

          total 4 drwxr-xr-x 2 root root 4096 Feb 16 23:10 fileserver

So it already exists, checking to see if its empty:

          ls /mnt/fileserver

This resulted in no return, there was no output to show. 

  mount -t cifs //192.168.50.200/sharedfolder /mnt/fileserver -o credentials=/etc/cifs-      creds,uid=1000,gid=1000

It couldn't connect, there was a mount error of 111. Ways to resolve can be to check to make sure the share folder exists, check firewall settings, and check network connectivity. Now I feel like I already did this, but because I feel like there are things about Linux I don't understamd, I am going to try and resolve the first step anyway. I got into a tangent where I was entering sudo commands and every time I did so, it wouldn't work so I went about trying to resolve this. I was going to enter:

        sudo pdbedit -L

I ran an update but got an error due to an attempt to access the Proxmox Enterprise repository which requires a subscription. To get around this, you edit the APT sources list. 

        nano /etc/apt/sources.list.d/pve-enterprise.list

Comment out the line that's already there and add another line (using # symbol)

        # deb https://enterprise.proxmox.com/debian/pve bookworm pve-enterprise

Add this line below it:

        deb http://download.proxmox.com/debian/pve bookworm pve-no-subscription

Went to update again. It still errored out, but managed to retrieve the necessary packages for the time being. I rn:

        apt install sudo

It installed this time. I went to add my user to the sudoers group (so I can use sudo). 

        usermod -aG sudo root

I was not familiar with what my user was called. I forgot that each line the curser is on tells you the user you're using. So I checked my SSH command at the beginning and it said that the IP address was a machine called root. But I ran the following command anyway to be sure.

        sudo whoami

The answer was root. going to run:

sudo mount -t cifs //192.168.50.142/sharedfolder /mnt/fileserver -o credentials=/etc/cifs-creds,uid=1000,gid=1000

The device at the address could not be found. So I pinged myself. The ping was successful. So I entered:

        systemctl status smbd

It returned:

        Unit smbd.service could not be found.

Samba is not installed. Now I used to just assume that SMB and Samba were the same thing but a conversation I had with Doug like nine months ago revealed that Samba is just a flavor of SMB. I thought Samba was another name for Secure Shell abbreviated so it was easier to remember or something. But I guess not. So in this case, Samba is not installed. 

        apt install samba

I think I successfully installed Samba. I ran the following command to verify it was installed and is running properly. 

        systemctl status smbd

It was successful and is running. 

mount -t cifs //192.168.50.142/sharedfolder /mnt/fileserver -o credentials=/etc/cifs-creds,uid=1000,gid=1000

Oh gosh how embarrassing. I just missed it, I needed to modify this last command with the information needed to get into the file server from the proxmox server. So I finally realized I needed to use the read only credentials I have for my file server. Okay, I went back to the nano file I think I created a while back with:

        sudo nano /etc/cifs-creds

This was already done. I ran the mount command to start the smb share between my file server and proxmox server. And I keep getting errors like:

        1000,vers=3.0

This is due to a formatting issue I think, and then:

        1000,vers=3.0

        mount error(16): Device or resource busy

        Refer to the mount.cifs(8) manual page (e.g. man mount.cifs) and kernel log messages (dmesg)

So it looks like the mount was already done so  I am checking to verify. 

    mount | grep /mnt/fileserver

It shows that it was mounted successfully. I chose to continue this through the GUI especially since I had so much trouble navigating my file server through the CLI because of the way I name folders. I logged back into Proxmox and went on the left side pane > clicked Datacenter > in next next side pane I selected Storage. However when I got there, only Local and Local-lvm were listed. So I started going about how to resolve this issue when I was wondering why the instructions were talking about CIFS instead of SMB. 

    My familiarity with CIFS from memorizing ports and their protocols is that SMB and CIFS both use port 445, and that CIFS is an older, legacy protocol. But the rabbit-hole doesn't end here. Proxmox prefers CIFS for redundancy, but as my file-server uses SMB 3.0, and CIFS is very similar to SMB 1.0, but Proxmox can still work with SMB 2 and 3.0, it'll just refer to them as CIFS. Proxmox will negotiate the highest available SMB version supported with both itself and the file server. 

    Using the GUI method of connecting Proxmox to my SMB server, I went Datacenter > Storage > Add, and this list was supposed to have the option CIFS to select. I had trouble finding it because it was listed under SMB/CIFS, then in the SMB/CIFS window I went: ID: Fileserver-READONLY, don't use spaces in this name! > Server [entered the file servers local IP address] > User: [the user you wish Proxmox to use to access the file server], which I have a read only user for > Password: [the password for that user on the file server] > Share: [the name of the share on your file server, which I have three, one for  me called KeepOut, one for my mom called MotherF because of a joke we have where her computer gave the hard drive I named Mother the drive letter F, and a third share where only certain things KeepOut has access to are available to be shared just in case someone gets access to me server that I didn't intend to have access. I would give them that share and read only so they can't delete anything. 

        I clicked Add. This did not work because Proxmox wanted to add a folder to my file server and I only gave it read only access. 

        So I am trying again in such a way where Proxmox will not need to make any changes to my file server and will only just retrieve files from it. I went through the whole process again thinking I needed to select Disk Image just for the same problem to happen again. And I realized I am supposed to select ISO image, and this dropdown menu is weird, anything highlighted on this menu will be filled on the space for the selection. I only want ISO image, so make sure when you select ISO image to highlight it for selection, and click Disk image so it is no longer highlighted. When the line only says ISO image, continue. 

    That had the same error. I determined that no matter what method through the GUI I use, it is going to want to add a directory to my file server, which I don't want because I don't want it to have write access. But it must create a folder for Proxmox, and so I thought about just creating a folder of whatever name Proxmox would name the directory on my file server so when it tries to access my file server and create a folder, it will see its already there and hopefully just connect to that. But then research found that it may try to change directory attributes which would again...require write access. I decided to switch to the CLI, my SSH window is still open. I entered the mount command again forgetting I already did it. And the error was that it was already created. I used the next command to remove it and another mount that was created in Proxmox to avoid confusion:

        sudo umount /mnt/fileserver

        sudo umount /mnt/pve/Fileserver-READONLY

I entered the mount command again the way I wanted it, and verfied it with the following command:

    ls /mnt/fileserver

This was successful, the shell listed files that are in my file server. Then I ran the following command to find the .ISO windows server file:

        ls /mnt/fileserver/*.iso

It said there was no .iso file. But the ls command a minute earlier showed the "'- 3 P r o g r a m s'" folder on my file server, and the iso is in another folder inside that folder. I showed this to ChatGPT because navigating these folders in linux with my naming scheme in windows is a nightmare. So I showed it to ChatGPT and ChatGPT caught right on saying that there are spaces in the folders name, so to CD or change directory to that folder I needed to enter the following command:

        cd /mnt/fileserver/-\ 3\ P\ r\ o\ g\ r\ a\ m\ s

It worked. Now I need to to the LS command, to see how linux displays the inner folder where the iso is located. It worked again. I did an LS command to show whats in the folder and there was the iso. I used the following command to copy the iso to proxmox for future use:

        sudo cp "/mnt/fileserver/- 3 P r o g r a m s/OS's/WindowsServer2025.iso" /var/lib/vz/template/iso/

I also used this opportunity to copy over the other ISOs in there so I don't have to do this again. So I copied each over one by one to Proxmox and verified them. Now it is finally time to set up windows server on a vm!

        Okay, Went to the Proxmox web interface, in the top right of the interface selected CREATE VM > under general tab of new window entered "WinServ2025" for my name of the VM > clicked next > In OS tab selected WINDOWS SERVER 2025 from the ISO image dropdown menu > for the OS TYPE or TYPE I selected Microsoft Windows > NEX > EFI STORAGE set to only option which was LOCAL-LVM > TPM STORAGE set to only option which was LOCAL-LVM > NEXT > under Disks for the BUS/DEVICE field I set it to SATA, leave the number 0 if you're using the default drive because this is to identify the drive being used > set DISK SIZE to 40 if just installing Windows Server with Active Directory and DNS and LDS and a domain controller, 60 to be safe if you plan on adding very many users or potential scalability, I went 60 for potential scalability > leave cache at NONE for simplicity unless you need performance. 

        In the CPU tab I thought about provisioning anywhere from one to three cores to this VM. ChatGPT suggested I allocate 2-4 depending on my needs. I expressed that I would not be running all VMs at one time but intend to install more than one, also having an Ubuntu Server VM as well as an Arch Linux VM just for practice and I wouldn't be adding many computers to the domain for Active Directory. I asked it what would happen if I assigned 2 cores to each and did have more than one running at the same time considering my server only has 4 cores to begin with, and no hyperthreading on this processor, and it said that the VMs would compete for resources, the system would get bogged down and be very load intensive on the processor. It would lag. ... For a second, I thought it was gonna be like that Simpsons episode where Bart meets a guy with a bursting appendix, and the doctor pulls out a pocket knife to cut it out and throws it, exploding like a grenade.

        I realized there's a problem, I was thinking of using a couple of laptops I have sitting around to be apart of the Active Directory so I can simulate an actual environment at least somewhat and then asking my instructor to break it somehow by giving him access to my VPN so he can do it remotely as a favor to me, but I realized, wait, AD is for professional environments, so only Windows Professional, Enterprise, and Education Editions can be used with Active Directory. Which means I would have to create a VM in Proxmox to install Windows with either one of those three editions, and it would have to run at the same time as the Windows Server VM with Active Directory or else AD can't manage it. So if the host OS which is Proxmox needs at least one core and everything else has to share three total cores, which gets what? ChatGPT suggested Windows Server with Active Directory gets 2 cores and Windows Professional or whatever I decided to use will use one core. 

        Okay, maybe I can get it to run a bit stripped down. At least this Proxmox server has 16 GB of RAM otherwise this just wouldn't work at all I don't think. And even so I predict there will be some sluggish behavior. But also, this is running of an NVME M.2 which I assume will also result in less trouble. I remember just how monumentally faster my laptop got ten years ago when I switched out the 2.5" magnetic drive for a SATA SSD. And NVME is even faster than that. I wonder if I should consider switching out the processor since that seems to kind of be the only real bottleneck here. 

       I set the Windows Server VM to 2 cores, and then I will create another VM for Windows Pro or something with 1 core when that comes up. On the Memory tab I set RAM to 8GB or 8192MB, and will set the windows VM to 4 GM since the system only has 16 and I want to save 4GB for Proxmox. In the Network tab I set Model to VirtIO (paravirtualized) > Bridge to VMBR0.

        I recall last time I set this up on VirtualBox on my laptop, there was an option to select SKIP UNATTENDED INSTALL. This caused quite the headache, and I had never seen it before and haven't seen it since. But I basically answered my question why just now because I have been wondering why I didn't see it any other time than when I installed it on the laptop and it was obviously because I was using VirtualBox and not Proxmox to run the VM. This caused quite the headache because I thought I understood instructions properly and no doubt because I think this option is poorly named, being called SKIP UNATTENDED INSTALL, it's a double-negative sort of thing and people have trouble with that. And I am struggling to keep straight if I want to skip it or not. It's not here in Proxmox so I don't have to worry about it but still! 

        I troubleshooted a connection error for a while, checking to make sure everything was fine in the temrinal, pinging the server and using trace route, everything was fine but I kept getting connection errors every time I would hit the finish button trying to create the VM, and so I retried creating it a few times, when I would click on the VM ID line a box would come up and that would say connection error. I tried:

        systemctl restart pveproxy

but this didn't solve it either. I clicked around and pages were now having trouble loading. I remember last time I was instructed to clear browser cache and this caused a host of other problems so since I am almost certain that's not it, I tried all the other things first. Finally I just entered the IP address with the port number separated by a colon in the search bar of the browser, and it didn't like that either. So I just totally closed the tab and opened a new one and it navigated fine, I entered my credentials fine, now it seems to be working as if nothing happened. So note, if you have a period of inactivity, proxmox will start to act like it's not connecting even though it is apparently connected for the most part. Just close the window and reenter the web interface. 

I re-entered all the info for the VM to be created and now the VM is created. 

I had a huge problem for two hours trying to get passed the error:

        BdsDxe: failed to start Boot0002 "UEFI QEMU " from /Pci (OxlF,Ox2) /Sata : Time out

        BdsDxe: failed to load Boot0002 "UEFI QEMU HARDDISK QH00013 " from /Pci /Pci /Pci 

        /Sata         Not Found

        >Start PHE over IPv4.

I kept trying to play with boot order in Options tab and the CD/DVD tab making sure the windows server iso was selected. I went into the configuration file using the commands:

        /etc/pve/qemu-server/

and

        nano /etc/pve/qemu-server/100.conf

Changed the line:

        efidisk0: local-lvm:vm-100-disk-0,efitype=4m,pre-enrolled-keys=1,size=4M

to:

        efidisk0: local-lvm:vm-100-disk-0,efitype=4m,pre-enrolled-keys=1,size=512M

Then I found a post that said "the iso is not signed, and can't be started using SecureBoot (which is the default on the latest OVMF shipped with Proxmox). If you hit the escape key while it says 'Startup boot options' (and before it says the UEFI message about saying hitting escape that doesn't actually work), then you get into the UEFI menu which you can use to disable Secure Boot and then boot the iso" and so I tried to play with the little side menu on button to press the escae key in the Proxmox console but I just couldn't get it rpessed fast enough before the error appeared. So I nano-ed into the configuration file in the shell again and added a line below boot order saying:

        boot: order=ide2;sata0;net0

        secureboot: no

It still didn't seem to fix anything, but I am hazy now if I turned the VM off again after this. I then tried removing the EFI line altogether so the config file went from this originally:

        bios: ovmf
        boot: order=ide2;sata0;net0
        cores: 2
        cpu: x86-64-v2-AES
        efidisk0: local-lvm:vm-100-disk-0,efitype=4m,pre-enrolled-keys=1,size=512M
        ide2: local:iso/WindowsServer2025.iso,media=cdrom,size=5873196K
        machine: pc-q35-9.0
        memory: 8192
        meta: creation-qemu=9.0.2,ctime=1740692522
        name: WinServ2025
        net0: virtio=BC:24:11:E4:F3:D4,bridge=vmbr0,firewall=1
        numa: 0
        ostype: win11
        sata0: local-lvm:vm-100-disk-1,size=50G
        scsihw: virtio-scsi-single
        smbios1: uuid=711b1975-1fa5-4927-a93b-9a1738dc0295
        sockets: 1
        tpmstate0: local-lvm:vm-100-disk-2,size=4M,version=v2.0
        vmgenid: f17cb149-5774-45d1-9bd9-4e964f671c55

To this:

        bios: ovmf

        boot: order=ide2;sata0;net0

        secureboot: no

        cores: 2

        cpu: x86-64-v2-AES

        ide2: local:iso/WindowsServer2025.iso,media=cdrom,size=5873196K

        machine: pc-q35-9.0

        memory: 8192

        meta: creation-qemu=9.0.2,ctime=1740692522

        name: WinServ2025

        net0: virtio=BC:24:11:E4:F3:D4,bridge=vmbr0,firewall=1

        numa: 0

        ostype: win11

        sata0: local-lvm:vm-100-disk-1,size=50G

        scsihw: virtio-scsi-single

        smbios1: uuid=711b1975-1fa5-4927-a93b-9a1738dc0295

        sockets: 1

        tpmstate0: local-lvm:vm-100-disk-2,size=4M,version=v2.0

        vmgenid: f17cb149-5774-45d1-9bd9-4e964f671c55

At some point while doing this and it still not working then going back to remove both the secure boot line and it had me remove the EFI line because I thought the configuration was broken and was going back to remove secure boot and add the EFI line back in, and then when refreshing Proxmox page and new web interface window and everything the start button was still grayed out, I entered the code to start the VM and it was already started so I went to the console and it had indeed started. The installation for windows server had started and was asking the language. I don't know what version of the configuration did it, because I had started to add the EFI line back and remove Secure Boot back out and it was working when I then went to add them back in thinking that fixed it but not being sure. I elect to not dig further into why this worked just to accidentally have it not work again. I will have to create another VM and try this again separately so I don't learn what happened just to not be able to fix it again and be back at square one. 

        I selected Windows Server Standard Desktop Experience > Accepted terms and drive configuration, took it a few seconds but it started installing. 

        I am working to try and get through this installation as fast as i can so I can stop or pause the VM and create another one and test the configuration file settings discussed above. 

        For future reference in case I can't find it I put WinServ2025 as the password to the login screen. 

        I have been unable to reproduce the total fluke that enabled me to install Windows Server 2025 on Proxmox. I have had friends suggest I redownload the iso, and also to just do 2019 instead of 2025. But I looked up the error again, realizing that when I googled it last time, the Power Toys text selection tool didn't perfectly reproduce the text when I pasted it so I searched it again and another page came up where they again explained the reason it wasn't working was because of a signature problem. But then a guy said that he just set the BIOS from (OVMF) UEFI to SeaBIOS, and it worked perfectly. So I have solace that I can at least reinstall the OS if I have to. But I still haven't figured out why the fluke happened. I created five different VMs ready to install just to try and figure it out and tried a different thing with each of them and only on the fifth one when I changed to SeaBIOS did I have success installing Windows Server again. 

This has been Truncat3d 00000000111100010100110______________end of line

2025-01-17 - Recovering Accidentally Deleted Files On RAID From Backup IronWolf Drive

    This last Friday, the 17th, I woke up with all sorts of errands in mind I wanted to get done because when I clocked out of work later that day, everything would be closed and there some things I wanted to take care of before that happened. But my supervisor had a personal situation happen that caused him to not attend work for over a week and I try to only wear my headphones when I can guarantee that I will not  get my headphone privileges taken away because I couldn't hear him. Since he was gone, I wanted to load up my phone with stuff to listen to, before getting on with the rest of my day. 

    I had a bunch of recordings of a family member who has passed on who had a really crass, irreverent and sarcastic sense of humor. I hadn't listened to them for a while and decided I wanted to hear them. But they are not named in a clean manner so they would all be organized at all. The naming schemes are all over the place or not at all. I wanted to clean this up when I saw it, so I put a few recordings on my phone from my Ubuntu RAID Server, which is minimized, meaning there is no GUI. But it uses SMB, so most of my interactions with it are through Windows, with a GUI. 

    I started organizing them with date at the front of the file names and then realized, you know, all it takes to lose all of these files, since it is Linux, is to accidentally press the delete key and accidentally confirm deleting these files. That is too uncomfortably close for comfort to risk. Right now what I do with several of my files which I peruse regularly is to put all of the files in a folder called "-0 actual" at the top of the folder in Windows, and then make shortcut copies of all of these files in the folder that contains the "-0 actual" folder. It doesn't eliminate the risk of accidentally deleting files, but because I will more often than not, only need to access the outer folder if playing files directly on my desktop computer. If I need the actual file I will have to go into the actual folder and get it and make sure to get out of it as quickly as possible to avoid fudging keys and stuff. I noticed that my mom fudges keys all the time which is why I removed write privileges from her. But I discovered that I also do it too, just not quite as frequently. I created users for myself that have no write privileges but SMB hates switching back and fourth between users all the time. So one of the tactics I employ is this actual folder solution. 

    As I moved the files into the actual folder I discovered that I had only copied instead of cut them and pasted them. So now there were two copies. And then I got mixed up over another problem I can't remember. I wasn't going to blindly just delete the files in the outer folder without first making sure I wasn't deleting anything that was not copied to the actual folder. And then I forgot which folder I was in and thought it was the outer folder and deleted all the contents. And then I realized what I had done. It is all very hard for me to understand how I did this trying be very careful, both folders were empty. It didn't matter to me how I did it. They were both empty and this was all that mattered. I started to panic. 

    I knew that the files should still be present on the IronWolf drive which backs up my RAID because Bo Landsem, a friend who has been a network engineer for twenty five years, told me that he never, ever under any circumstance whatsoever relies on a RAID, he always, always, always backs it up and that is why when all of his colleagues have RAIDs fail on them and scramble to try and recovery the data with special software and stuff like that, he always just reverts back to his other backups. So on his advice, I bought a 12TB IronWolf Hard Drive on discount while building my server, deciding not to place my trust in just a RAID. 

    At the time, Doug thought that the RAID was perfectly trustworthy but I refused to risk it. However, I did not know at that time that in Linux, all it takes is to press the delete key and confirm it before the file is gone. In Windows there is at least a recycle bin. Doug said things like, well, just don't press the delete key. No, I know that I will do it on accident. I know I am not perfect, so I need fail safes. 

    And what Doug explained like six months later when this information came to my attention was that there isn't really a recycle bin in Windows as we like to think of it, the files don't get transferred to a new folder, what happens in Windows which we see as a recycle bin folder is that windows flags the files so that they appear in the recycle bin folder but they don't get deleted until they have either been in there long enough to be automatically deleted or you clear the recycle bin yourself. I had made a habit out of always keeping my recycle bin clear and things like this have really caused me to want to be more careful and not do so unless I am sure I know what I am deleting. But in Linux, none of this matters, it's all just academic. 

    I was very uncomfortable with this. But I tried to fix it and make it so things can't just get deleted so easily, and I have not found a solution that satisfies. Seriously, in Linux, the OS that is by far the most customizable? And I can't get a solution that buffers potential deleting like a recycle bin?

    It just occurred to me that surely someone has created software for Linux that does this. Surely I am not the only one who thinks this. So I am going to look. But right now, I have accidentally deleted a bunch of files that have serious sentimental value to me and they cannot just be replaced. They are fortunately supposed to still be on my IronWolf drive for reasons exactly like this. That was a well spent nearly $400, I think on discount from about 400 for like 300 or something when I bought it. 

    But I had never navigated to it without Doug's help. We set it up and he navigated to it and showed me it was there, and then several months later he showed me how to navigate to it and that it wasn't that big of a deal, but that wasn't something I was going to remember. I figured, well hopefully I still either have his help or ChatGPT can explain how to do it or something. For now, I worry about what's currently on my plate and not about things that are not foremost on my mind. Someday I will better figure out how to do this, but I know it is there and that it is doing its job and I know that there is a cron job in Linux running an rsync every Wednesday early in the morning while I am asleep. And since  just barely accidentally deleted these files and it's a Wednesday, I don't need to worry about whether or not they are there. But my whole day just went out the window, I will call in sick if I have to in order to avoid losing these files. I need to know they are recovered so I can peacefully move on with my life. And I have several hours before I need to be at work so there's a good chance it won't come to that and I need to make sure I don't have to call in sick because I don't like short paychecks. 

    First thing, I opened up ChatGPT and asked how to navigate to this file from the directory where I SSH in from. To my surprise this way far easier than I thought it would be. And I thought, once I get to these files and see them listed on the cli, I will SCP them over to my laptop. I don't want to just SCP them back to the RAID, want to have them more tangibly in my hands in a fashion I know that I have them and know what I am doing and that by dragging and dropping them in the folder of my choice, I am not mistyping a command or some stupid thing and anything happens that with my current comprehension, is unpredictable. So I wanted to SCP them to my laptop which is where I was SSHing into the server from. But I had a bunch of challenges that required troubleshooting and caused a great deal of stress. 

    Just so you know, I am leaving out much of the two and a half hours of frustration not figuring out why things weren't working, but I wasn't able to simply enter the folders I desired. I first had to contend with why I couldn't see the files I was expecting to see. I don't remember why I was able to look into these files so easily and see that what I was looking for was missing, but suffice it to say, that I saw firstly that there were a bunch of users such as "KeepOut" and also "KeepOut[with a skull and crossbones emoji]", Jack Sticka's idea. After dealing with that for maybe nine months and having had to set up SMB several times and having to navigate to find the proper emoji online and then deal wuth the fact that the same exact emoji looks different on one ecosystem such as apple than on another like windows. I would search on my iPhone and not see exactly what I saw in Windows. I ultimately discovered that Windows and Apple display the same exact things verbatim in totally different ways. So I eventually got Doug's help to remove the emoji from the user. But on the IronWolf, this was not straight forward like it would be in Windows. The same user put the same files and folders together, but it was named slightly differently when it did it in a number of cases. I entered "KeepOut" thinking that everything will be in there. I don't know why the version of the user with the emoji is there, but when I went inside to look for the folders and files, the folders were there but the files were not. I backed out eventually and tried the user with the skull and crossbones. And then for some reason, I then experienced great difficulty in actually accessing these files. Now this difficulty only lasted for maybe a half hour. What came next was the real difficulty and it compounded later. 

    There was a folder that in Windows, is called "- 1  Audio", and then inside of that, in Windows, is called "dad A U D I O". This is how I control where the folders are regardless of word the file is named with, if it is a commonly accessed file, I want it near the top and not have t scroll through and searching with a fine tooth comb through folde3rs that almost never get accessed. I prioritize the ones that I access all the time with special names and characters and extra spaces. And in the cases where I simply have a lot of folders and need to be able to find what I am looking for without visual strain, I is capitalization and spaces throughout the name. 

    So "- 1  Audio" is for organization and is the outer-most folder I need to access. And "dad A U D I O" is the inner most folder I need to access in this case and is for visually standing out in the crowd so I can find it easily. This caused so much trouble. When Doug was helping me set it up, I had unwittingly done this in Windows, not knowing that this was a problem with Linux. So when we had to set up permissions for different users, these special naming schemes which don't appear too complicated in Windows suddenly got extremely complicated in Linux. And it isn't even that what you have to do in Linux is so complicated as a result, but keeping track of what you are doing so you do it properly is a nightmare. 

    In Linux, you have to do something called Escaping spaces with a backslash. For example: A file named my file.txt is referred to as my\ file.txt.

    You can see how just one space may not seem complicated but how with "- 1  Audio" and "dad A U D I O", this is actually far more complicated: "\- 1\ \ Audio" and "dad\ A\ U\ D\ I\ O". ChatGPT chooses now to let me know that you can simply type commands like cd to change directory, directories in Linux being the same as folders in windows, using the cd command and the name of a directory can either be done this way, escaping the spaces with backslashes, or you can type the names of these files with quotes. So instead of:

    cd \- 1\ \ Audio

and

    cd dad\ A\ U\ D\ I\ O

...you get:

    cd "- 1  Audio"

or

    cd '- 1  Audio'

and

    cd "dad A U D I O"

and

    cd 'dad A U D I O'

    I sure wish I had known this because that would have made accessing these files and then doing SCP later, a million times easier. Can I exaggerate in this case? It was extremely hard and frustrating. I kept not knowing what I was doing wrong. I enlisted ChatGPTs help and to make my not knowing what I was doing worse, ChatGPT chose this moment to get everything wrong. I hindsight I would have specifically googled how an SCP command is supposed to look so I could just verify that ChatGPT was doing it right whenever I turned to it, but this is the problem with panicking. I can at least say in this case that if the files are not my files, I won't have to panic and can focus more. But unfortunately, I am too familiar with how these things go than to think it's that simple. There will always be something that can potentially be panicked over, like you accidentally deleted crucial files for your job and you will get fired if they aren't recovered. 

After successfully navigating to the correct directory on the Linux server, I attempted to use the SCP command to transfer files to the Windows machine. The initial SCP commands failed due to several reasons, including incorrect IP addresses, once I had the proper IP addresses, it took tie to then realize I was getting them backwards, and that the IP for my laptop was used for the server and the IP address for the server was used for my laptop. Then there were improper file paths, and incorrect user credentials. I verified the IP addresses of both the server and the Windows machine and ensured the username for the Windows machine was correct by checking the prompt in the Windows terminal, which displayed the logged-in username. And I used IPCONFIG in windows to verify my windows laptop IP address. I don't recall using IFCONFIG to get the IP address of my server...oh yeah, that is because I have SSHed into my server so many hundreds of times by this point and the address is reserved on my routers DHCP address reservation configuration page, that I didn't need to look it up. But an IP address reservation is different from a static IP but gets the same result, just a static address is configured directly on the machine and a reserved address is always the same address given to the machine by the server. I am not sure why if you statically assign an address, why you wouldn't still have to reserve the same address on the DHCP server, because you could see that if the DHCP is dolling out different addresses than the computer has statically assigned to it, then network communication will cease unless maybe broadcast messages maybe get through. And you need your printers and servers to always have the same IP address or things like file shares would be a nightmare and would never work. 

    The next hurdle involved the SCP command structure. On several occasions, the command failed because it did not accurately reflect the directory structure or handle spaces in file names correctly. Commands such as scp -r /mnt/ironwolf/dad_audio/* Christian@192.168.50.65:/C:/Users/Christian/Desktop/Dad_Audio_Transfer/ were adjusted multiple times to address issues like inaccessible directories or improperly defined paths.

    ChatGPT suspected network issues when every time I entered a command and the text curser on a blank line under the entered command would just blink endlessly, sometimes for five minutes, sometimes ten and sometimes a half hour, before finally giving up. I confirmed that the SSH connection was still active by successfully navigating and executing other commands on the server. Despite this, SCP repeatedly failed to establish a transfer connection, which highlighted syntax issues in the command. 

    To simplify the process, I explored alternatives to SCP, including using rsync. However, this also presented challenges, such as ensuring rsync was installed and properly configured. Ultimately, I returned to the SCP command, but after checking to make sure the correct Is were used, then swapping them so the destination wasn't the source and the source wasn't the destination, then making sure the usernames on both the source and the destination were correct, I still didn't get SCP to work. These all contributed but were still not what the problem was. I never figured out why the SCP command didn't work ultimately, but what allowed it to work at last was that I had opened WSL in another Terminal tab, and was exploring the possibility of a few things there but somehow by chance, I thought to try using the SCP command from WSL on my laptop to reach over to the server and grab the files there and pull them over to the windows machine, where I then used SMB to put them back on the RAID. 

I believe the final revision of the SCP command was:

    scp -r * Christian@192.168.50.65:/C/Users/Christian/Desktop/Dad_Audio_Transfer/

    -R means recursive, any file in said folder that this command points to, as well as any folder inside the folder this command points to, will also be copied. But the asterisk is a wildcard command. In this case, it tells the computer that this file we are currently viewing, this directory, plus every folder inside it because of the -r command, will be copied over. Usually the SCP command is laid out in the fashion of "scp" followed by options which in this case is "-r" and "*", both typed separately from each other, then the username@source host Ip address, the source file path of the files to be transferred, then the username@destination IP address, with all three of these separated by colons and no spaces between the locations. But the file path of the files to be transferred was indicated in the options section of the command with the asterisk, so the file path or name of the folder the files were in was unnecessary. And it should be understood that if there are spaces in file or directory names, those are escaped out with a backslash before the space in each case, or the whole title is quoted. There are no spaces between the three addresses, the user@source ip:directory name if it's local or the file path if it's local or not:and the user@destination ip. 

    In conclusion, resolving this issue required iterative troubleshooting of directory access, user credentials and IP addresses, and SCP command syntax. 

This has been Truncat3d 00000000111100010100110______________end of line