Hi can someone tell me if I would be making a mistake to build a Thread Ripper Pro Linux box to run LXD for desktop VM’s ?
Also thinking about building a Epyc server for Linux back end stuff. any tips or opinions would helpful
1 Like
I don’t know why you would. Linux can be run and run well on almost anything. I’m currently running a non-pro Threadripper system with Arch installed and and while it wasn’t perfect the issues I had were software related not hardware.
Hi TeamTux,
I am planning to build a LXD server to run workstations (heavy loads) in containers. Thread Ripper 64 cores give me a lot of containers to work with
I understand and Threadripper can be a good fit for that. But do remember application should drive hardware choice. If I were to setup a dedicated server I run Eypc and for a HEDT or workstation I go Threadripper. Each of these products fits in market in different places despite those sector each being rather nuanced. That is just my two cents and were I looking at that scenario I’d be looking for hardware that is very fitting for my application rather then trying to fit a square peg in a round hole.
Hi TeamTux,
So what I am planning to do is run all my workstations VM’s in one LXD server. they will be pen testing, test labs for Linux workstations and other workstation related stuff like development.
1 Like
In my limited experience with TR 1950X, there were a LOT of teething issues, present to this date. I have given up upgrading the UEFI as the systems have run fine as a Win 10 Pro boxes, my main one is my primary HTPC rig.
My most stable VM platform is a Dell R740xd and I just added 2x WD Gold 16TB drives as a second RAID1 array. The Dell units are really impressive; mine’s a bit older (about 2-3 years on since getting it) with 5-year warranty and 4-hr mission critical warranty. 32c/64T dual Xeon + 256GB RAM and a plethora of copper and fibre 10G (Intel nics FTW!).
Intel hardware has always been the most stable for me with XCP-ng (my go to hypervisor, simply as I haven’t had time to try Proxmox).
I’m running a dual Super micro cluster for my Gitlab server but I’m having issues getting a second SSD being added a SR (even via Xen cli). So I’ve given up on that for now; I think I’m missing some BIOS config in the SuperMicro. For now, the existing boot SSDs are enough to run Gitlab and a few “slow” workers; I’m running much faster workers on the Dell for my Rust compilation and docker builds.
Oh another tip, take the rack enclosure approach and make sure you get a 1000m (1M) deep rack at the start, this is needed to support the Dell R7xx range of units but do you homework as I’m sure there are deeper variants.
Hope this helps.
1 Like
What are you running on LXD though?
This topic was automatically closed 273 days after the last reply. New replies are no longer allowed.