TrueNAS

UGREEN DXP8800 Plus – TrueNAS 64TB SSD

It was finally time to make the living room more silent, by replaing my old TrueNAS Core (based on FreeBSD) with the newer TrueNAS Scale (based on Debian Linux).

Instead of building it myself again, like my old
SUPERMICRO A2SDi-4C-HLN4F with 32TB on Raid-Z1,
I decided to buy a complete Hardware-NAS-Chassis with 10G Ethernet.
After some studying, I decided for the previous Kickstarter NAS project UGREEN NAS DXP8800 Plus, which allowed me to swap the firmware back to my beloved TrueNAS.

All the harddisks had to go, I want solid state only, for power saving and mostly for accoustic reasons.
So, I bought 8x 2,5 inch SA500 NAS SSD’s from Western Digital for the 8 slots, still with slow SATA interface, and built a RAID-Z2 this time, for long term archives.

And for fast and temporary data like AI/ML LLM+Diffusion models, backups and software archives, I bought 4x 8TB Western Digital SN850x NVMe SSD’s.
To be able to use 4 M2 slots, I bought an extra PCIe adapter with bifurcation capability.

I did also replace the original fans with more silent Noctua NF-A12x25 fans.

And to round things up, a 10GBit/s switch needed to be used with my MacStudios, so they have lightning fast access to the LLMs 😉
An SFP+ version might have been cheaper, but I decided to use the YuanLey YS100-0800T with 8x 10G/5G/2.5G/1G ports, so I can also get more speed for my Proxmox nodes which currently do 2x1G bonding, but are 2.5G capable. Plus my Ryzen Windows machine has a 5G port.

Of course memory is maxed out with 2x32G, although I’ll not use it for containers anymore, only for ARC cache. My MacStudio Podmans are good enough for that, and have faster access to their local TB5 storage enclosures.

Food

A Taste of Thailand at Mun Mun Restaurant in Munich

Our family had a fantastic meal at Mun Mun in Munich! We tried two popular Thai dishes:

  • Pad Kra Pao: Ground beef with holy basil, spicy and full of flavor. Super delicious!
  • Pad Thai Noodles: Sweet, sour and packed with chicken and veggies. Authentic taste straight from Thailand.

The staff is super fast and super friendly, and all the food is so authentic.
But be aware the portions are huge here.

MunMun
Software, Wordpress

Blog site moved

I finally migrated my wordpress blog over from https://teddy.spacetechnology.net to https://myblog.kobaan.nl

The whole thing now runs in a podman pod on the “smaller” 256GB MacStudio, as compared to the tight FreeBSD jail before.

All my docker images are automatically updated using watchtower, and wordpress also uses auto update.

Unfortunately the migration broke 5 times, as I had to readjust resources a lot. And finally changing the domain broke 1400 links, which I had to repair. Now still some images are mysteriously unlinked although the media browser can see them. I try to fix that the next days…

LLM

LLM – Optical problem solving now works

Finally got the PDF OCR flow working, and the LLM can now solve math problems. Only grey diagrams are still hard for it to recognize properly.

I used a math training exam that my son is currently working on in preparation for Gymnasium.

And even the 27B model is sufficient to solve this level of math.

Apple, Hardware

Yet another Mac Studio M3 Ultra this time with 512GB

Did I say enough…..
meh….

Okay lesson learned, never enough.
My “1 month old” M3 Ultra 256GB went out of memory running all those models and podman containers in parallel.



My current setup is:
OpenWebUI:
-> LM Studio, Aya-Vision-32b,
-> ComfyUI workflow with t5xxl, Flux1.dev, llama-3.1
-> MCP proxified: Searxng, wikipedia, docling, context7, time, memory, weather, sequential-thinking
Podman: 24 containers including supabase, wikijs, watchtower…

Also I discovered that I can use OpenWebUI, SwarmUI, exo and even mlx
to distribute workload across both Mac Studios via 80GBs thunderbolt 5 bridging.

And with the orange clown, you never know if there will be a new M4 Ultra next year at all.

LLM

LLM full throttle on M3 Ultra 84 Watts !

What a crazy efficiency monster.
Running a LLM on all 80 GPU cores and the system is only drawing 84 Watts…..

Nvidia must be crying at night.

and this is only M3, M4 is even more efficient, but yet not available as an ultra fusion variant, and end of year Apple will manufacture in 2nm