This sounds really cool. Do you have any guides or notes on how to set this up? I've been looking for something to dive into for LLM and MCP.

Would be interested in this too!

+1

You can also ask for a remake with no questions asked.

To me, their coffee is consistently bad, so I have to assume that what I got is the way they wanted to make it. A remake will not change things.

Until writing my comment, I never realized people thought Starbucks coffee is bad.

Where can one get good coffee? I want to try!

Does anyone have a good _basic_ guide on LVM/LVM Thin? I'm having a hard time wrapping my head around LVM and moving the vmdk to it. Mainly a Window admin with some Linux experience.

I understand that LVM holds data in it but when I make a Windows VM in proxmox it stores the data in a LVM partition(?) as opposed to ESXi or Hyper-V making a VHD or VMDK.

Kinda confusing .

proxmox is using LVM for direct attached raw volumes. LVM is just a logical volume manager for linux, which gives you more features than using old fashioned disk partitioning. I guess they chose this path for windows virtual machine migration because windows running on vmware before, does usually not have the required virtio drivers installed to support the qemu hypervisors virtio solution for disk bus virtualization out of the box. It would mean the hypervisor has to simulate IDE or SCSI bus which comes with great overhead perfomance wise (in the case of migration)

So an direct attached lvm volume is the best solution performance wise. In the vmware world this would be an direct attached raw device either from local disk or SAN.

For fresh install on proxmox its better to chose qcow as disk image format with virtio-scsi bus (comparable to vhdx, vmdk, qemus disk format) and add virtio drivers during windows setup.

i ran into the same sort of documentation desert.

off the top of my head:

- keep in mind there is LVM and LVM2, and proxmox now uses lvm2

- I don't understand the thinpool allocation. You don't have to use lvm-thin if you don't want to deal with oversubscribed volumes, or don't care about snapshots or cloning storage.

- get to know "pvesm". A lot of things you can do in the gui

- when making linux VMs, I found it easier to use separate devices for the efi partition and the linux partition, such as:

  efidisk0: local-lvm:vm-205-disk-0,size=4M
  virtio0: local-lvm:vm-205-disk-1,iothread=1,size=1G
  virtio1: local-lvm:vm-205-disk-2,cache=writeback,discard=on,iothread=1,size=32G

(virtio0 = efi, virtio1 = /)

and I can mount/expand/resize /dev/mapper/big2-vm--205--disk--2 without having to deal with disk partitions

I remember the same issues but don't know why we had to enable IPX/SPX over TCP/IP. I have a specific memory of troubleshooting for an hour and then clicked IPS/SPX and StarCraft worked.

Does anyone remember a specific reason why? I was too young to know or care.

For a while, before TCP/IP really became "the" network protocol, IPX/SPX was also around and a lot of games used it. Red Alert 2, Descent 1 & 2, etc all used IPX/SPX.

Yeah, setting up IPX circa 2010 to play Dungeon Keeper 1 in multiplayer 15 years late was quite the puzzle ! Especially since I didn't have admin access to those computers... at first.

Can you share a link to the model that you're running?

They have 12k feet of cables?

No, they have cables over twice that long...

https://www.whoi.edu/what-we-do/explore/ships/ships-atlantis...

Worked on a software system for deep sea cranes, where they wanted us to be able to support them having 4000-8000m cables on the crane.

Any links to an ok $1 VPS that would work? I tried a few a while back and always have an issue with RAM.

What's the best way to go about clusters since I think you need 1gb RAM per node and lowest I've seen is $5.

Thanks

Thanks for this.

That's exactly what I want to do. Do you have any guides you used? I'm having a little trouble with nfs and docker permissions (guid/uid I think) trickling down to the nfs share in my Synology.

With regards to NFS permissions, I set my Synology up as so:

* Client: 192.168.0.8/29 (whatever the smallest subnet is I can use that includes my NUC)

* Privilege: Read/Write

* Squash: No mapping

* Asynchronous: Yes

* Non-privileged port: Denied

* Cross-mount: Denied

This is fairly out of date (I keep my own stuff in a private monorepo, this was a snapshot I posted for a friend), but here is the basics of a script to setup my NUC: https://github.com/regner/homelab-stacks/blob/master/scripts...

Of note, the "Adding NFS share to fstab" part doesn't actually do that...

Thanks, looks very interesting.