I tested SMAPIv1 on XCP 8.2.1 against SMAPIv3 on XCP 8.3b2 using the same host (a HP ProDesk 400 G6 with a i5-10500T CPU, 32GB RAM). A 1 TB Samsung 860 EVO SSD drive was used as the test SR, while XCP was booted from a 512 M.2 KIOXA NVMe drive. Fio (fio-3.37) was compiled from source on an up-to-date Debian 12 VM (2 vCPU, 4 GiB RAM, 32GiB drive) which was copied twice so that three identical VM could run fio in parallel.

After an initial fio run to create the files, a script run three sequential write and read tests (e.g. fio --name=fio --ioengine=libaio --randrepeat=1 --direct=1 --fallocate=none --ramp_time=10 --size=4G --iodepth=64 --loops=50 --group_reporting --numjobs=1 --rw=write --bs=1M). The script first ran on one VM, followed by a run on three VMs in parallel. IOPs and bandwidths were averaged.

d4beab9b-1328-4d67-8794-49b45093572b-grafik.png

v1-1VM are the results for one VM on a SMAPIv1 SR (XCP 8.2.1) while v3-3VM are the results for three VMs in parallel an a SMAPIv3 SR (XCP 8.3b2).

While I'm not sure if this approach is really valid (e.g. the average load of the host went through the roof when three VMs performed fio in parallel), it does suggest that the bandwidth of SMAPIv3 is not yet en-par to that of SMAPIv1. But I could be wrong and this is an early previews of SMAPIv3. Looking forward to more performance results on SMAPIv3.

@shinuza oh, yeah. they've deprecated it in the last couple of months. thanks for the heads up

@thepro101

i would install the guest-tools right from the start after base install of the os.

create min 3 nodes, install them as you like with guest-tools and then install the first rke2-node. read the docs how to config metallb on rke2. you have to do a small config tweak in the nginx ingress controller config

then add the other nodes.

to manage rke2 i would use openlens.

you can use democtratic-csi but be aware that you completely trust truenas and this opensource-project for your data. i dont think that i would go this route when i am not expirienced with k8s.

i would use the nfs-provisioner and when everything works fine and you have a solid csi-enabled backup you can add democratic-csi in the mix.

for us, backup and restore was the biggest problem. in theory everything seebs easy with k10 or velero but if you completely shoot your cluster you will have a very hard time.

to be honest, after 6 month and some installations we gave up on k8s and migrated our customers and our internal IT to a setup where we use openSuse MicroOS VMs for every docker-compose project. We now have approximately the same amount of VMs which we had as namespace but with the benefit of complete control over resources with very little overhead. And we have the benefit of an optimal backup and restore.

K8S bite me quite a bit too often 😉

Tools is working on Ubuntu 24 LTS.
Well it obvious, since nothing is really changed)

@Nux rebuilding the dom0 kernel is quite tricky, and I am not sure it is going to help your usecase, if you really want, you may want to look on xcp-ng-rpms and the related linux rpm
Regarding nested Hyper-V in Xen, gen2 support was considered in the past, but it never got fully implemented, see https://lore.kernel.org/xen-devel/20191130115737.15752-1-liuwe@microsoft.com/
and https://lore.kernel.org/xen-devel/20190923100931.29670-1-liuwe@microsoft.com/ for related discussion on the Xen mailing list.
So as it is now, Xen in Hyper-V gen2 vm is not supported.
Only gen1 is supported on a best-effort basis.

Gen1 use realistic-behaving emulated devices, thus there is no Hyper-V specific driver that may improve performance.

@jasonnix depends on what PV driver you want to implement.
But such place would be the Xen MiniOS repo (tiny PV kernel project that serves as a base for Unikraft among others) : http://xenbits.xen.org/gitweb/?p=mini-os.git;a=tree

Also give a look at https://wiki.xenproject.org/wiki/Unikernels

@gb-123 said in USB Passthrough Override Script - to ensure usb-policy.conf consistency:

This is assuming pushing a custom file through XAPI is possible .....

That's the whole point : it isn't until we code support for it in XAPI.

@psafont said in Stunnel - Future plans to use something else?:

@nikade We've already fixed some issues to start using ocaml 5 regarding the C interfaces. Handling threading in ocaml 5 is still an open problem that the ecosystem has not yet solved (there are many libraries competing now). We still need to create a credible strategy to port xapi to the new model, and don't have any timelines yet

Yea porting it is probably not a bad idea, but I guess that involves the biggest tasks as well 🙂
All tho if you port it there might be opportunities to re-do a lot of things that you have been looking to fix for ages.

Crossing my fingers that this is something that you get time to do in the near future, there is a real chance here to take a big part of the esxi market as a lot of ppl are migrating.

@olivierlambert I think so. ceph.ko anyway isn't a core module for XCP-ng, so I don't think there's a high risk in patching it, especially with patches coming from upstream kernel.org.

Regarding the initial question, XCP-ng 8.3 will also have a newer kernel-alt. However, I don't recommend it for production, because it is a lot less tested in the context of XCP-ng.

@TheiLLeniumStudios I'm trying to do something similar with Talos, were you able to find a workflow that worked for you since this post?

Hi,

It depends on your Windows version (eg > 2019 : go for XenServer drivers). Those PV drivers are pretty stable and works. Alternatively, you can use https://www.xenserver.com/downloads

Our latest update train should provide full vTPM support or close to that (even for migration/snapshot). It's 8.3 beta 2. See https://xcp-ng.org/forum/post/68789

Yes, we are considering OpenELA for many reasons. It's backed by big companies with a common goal, so it's probably less risky than other choices (but Alma is also an interesting choice anyway).

As usual, it will be loosely based too: it's relatively impossible to "transform" a regular distro to XCP-ng, but it is very helpful to re-use many packages as possible for reducing the burden of maintaining them all 🙂

Well, it's not really possible to answer with a short answer. Globally, for the whole platform, we are talking about hard in a way that it takes time.

@ThierryC01 said in XCP-ng Windows PV Drivers 8.2.2:

Why is there no video drivers for Windows, I have a Windows 10 Pro VM and the video lags a bit compared to my Linux Desktop VMs despite having 8GO RAM and 16MO (?16 max???) video memory.

Because it's rudimentary GFX and has no acceleration at all. It's not made for what you might expect.

Hi,

No, nothing urgent to update them. It's mostly minor changes 🙂

@reinvtv @jedimarcus Checking Netbox feature requests right now 🙂 Would this feature still be important to you?

I'm also wondering: at the moment, there's a button in the pool's advanced tab to manually run a synchronization for that pool. But with multiple instances, how would that work? Would it synchronize the pool with all the configured instances? What if you only ever wanted to sync pool A with instance A and pool B with instance B?

@Andrew said in RTL8153 Compile:

@etomm For the code I compiled it mostly as-it (with one change to support the different XCP kernel header version). I also added three udev rule changes. One is included with the driver code. One change to stop XCP scripts changing USB ethernet names. And one that is to support static USB ethernet names and rename unknown USB ethernet devices to protect the system from conflicts.

I now have it working correctly with multiple USB ethernet adapters and stabled names that don't interfere with regular ethernet adapters. It still requires manually adding the MAC addresses and ethernet device names because XCP/Xen does not support USB ethernet devices.

@Andrew would you be willing to share the udev rule changes that you made? Also, are you using 8.2 or 8.3 beta? Thank you!