Search results

As I wanted to learn working with RDMA, I got myself a ConnectX-5 (MCX512A-ACAT; 2x25GBe); a second one will follow soon. Now, in addition to doing RDMA (and zero-touch RoCe in particular), these chips also have extended eSwitch functionality that is exposed using the Linux switchdev module. I...

I believe that is supposed to be called NOTWORK.

"Yes"

Angry European noises I wish we had outlets like that over here; I would have been in for a few of those.

These switches (along with their 4 x QSFP28 variants) do not support the 802.1 bits for RDMA, however, the ConnectX 4 and up do support something called zero-touch RoCe that should work even on not fully compliant switches. Has anyone had success using zero-touch RoCe on the Mellanox/NVidia...

This is good news indeed. Do be aware of the fact that your network-interface names in the kernel may change, so that you definitely need physical or IPMI access to your server when upgrading. See the discussion on this forum here.

FWIW, using the mechanism in https://pve.proxmox.com/pve-docs/pve-admin-guide.html#network_override_device_names worked well for me on another server.

I just upgraded two of my servers from Proxmox VE 8.1 to 8.2, and although there is just of good stuff, as outlined in the video , the upgrade to kernel 6.8 (although good in itself) changes the network-interface names in the kernel. In my case, on an Intel X710-DA2, the names changed: From...

That is one way to go. Personally, I have Caddy handle the SSL certificates and HTTPS work. Then, I use Authelia to only have authorized people (me) be able to even access the IPMI web pages.

"Lies, damn lies, and benchmarks" Whenever one defines success metrics, some will game the system for the sole purpose of achieving said metric. Another way of putting that is: "Any metrics that starts being used as a performance evaluation stops being a useful metric at precisely that time."...

I understand that this is Asus' 14th generation of product, but man, would SI3NA have been a great match, name wise :)

FWIW, Proxmox VE started including a fixed 6.5 kernel with a ZFS 2.2.2 module in it's pvetest repository yesterday, so the ZFS bug is now mitigated.

Thank you. I actually do mean the X13SAZ-F though, specifically a board with IPMI. What I was trying to find out is whether one can use the Intel IGP (for, say, Plex) when also using the IPMI for remote management, which is one of the use cases that I would like to have.

Is there anyone here with a X13SAZ-F who would be willing to post an lspci output here please?

AVX512 is "nice to have".

In a community colo, I have been running my Supermicro-X10 based Haswell server for a long time now, and it has turned 9 years old. I would like to refresh it, and I have been weighing the pros and cons of whether to go for Supermicro X12 or X13, especially since I also want to replace the...

This is fascinating stuff. In the Web/Tech industry, we are by now much more used to scale-out solutions (and our workloads scale well on those), but there are certainly workloads that benefit from the massive with-system bandwidth that you get in a scale-up solution. What does surprise me is...

That would indeed be interesting: PCIe 2.0 x1 is 500 Mbyte/s full duplex, so a 2.5 Gbe switch would be a feasible improvement. 5 Gbe would be pushing it, and 10 Gbe is more than the link can saturate. Also, in my case, Ceph is not very friendly on these small CPUs. Interestingly, in contrast...

Build’s Name: Malina (meaning "Raspberry" in various Slavic languages) Operating System/ Storage Platform: Plain Debian 11 plus LXD and Ceph CPU: 6x ARM Cortex-A72 Motherboard: DeskPi Super6C Chassis: SuperMicro SuperChassis 510T-203B Drives: 6x WD Red SN700 (1000 GB, M.2 2280) RAM: 6x 8GB...

And how about the reverse: Does this motherboard allow one to use a processor with a GPU, and then use both the IPMI and the GPU at the same time (e.g., for Plex pass-through), such as for example on many of the X**-**H-F motherboards?