xrstokes

Doh! Thanks for the quick response though.

I've been trying out UnRAID at home with the plan of taking it to work. All looked great but when I tried it at work it doesn't work with the 9361 8i we've got. I don't want to stop using it either because it's connected to 8 disks with some fancy expensive cachecade 2.0 going on. The performance of the raid card is too much to loose. Have I buggered up somewhere or is this just not possible? Thanks in advance.

Wow, I never even considered how the memory channels could play into all of this. Thanks so much for the detailed response. 👍

I've sorted it for now by flushing the disk cache before launching the VM. Then it get's it all from the second node.

I was just looking for this info, I noticed so bad RAM performance on my dual xeon server. Running numactl --hardware and shutting the machine down showed all the ram was allocated from numa node 0 when all my cpu cores were pinned to numa node 1. I added... <numatune> <memory mode='strict' nodeset='1'/> </numatune> and it almost worked, it took all the RAM from node 1 down to 2.5GB and then pinched the remaining 6GB from node 0, suppose that's better then not starting at all. Since my main vm uses the entire 2 CPU, [I have isolated that cpu from unraid]. Can I persuade unRAID to keep away from that ram? It seems to use it up pretty fast from boot.

Thanks this helped me a lot. Made a big difference to get it like that. Now just to set up static huge pages on my second numa node.

I second this, I have a dell t620 with 128G of RAM and I a guest that is pinned to the second CPU. I want all of that guests 64G RAM to come from the second numa node and all my other vms can just use THP and be on the first core with unraid. THP is quite cpu intensive. Would something like this work. https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/performance_tuning_guide/sect-red_hat_enterprise_linux-performance_tuning_guide-memory-configuring-huge-pages Thanks in advance, only been here a couple of days.