Has anyone setup a gen11 ProLiant with XCP-ng?
We are migrating from vmware to XCP this summer and will be replacing our current gen10 proliants with gen11’s.
I have been messing with a gen9 and two gen10’s to test out XCP and XOA so far no issues, just curious if anyone has ran it on a gen11 yet.
Not sure if this helps, but I’m running my lab on DL360p gen 8 and everything seems to work fine. Since you are several generations newer, shouldn’t be a problem. I also have some Supermicro with Xeon Scalable for my production system, no issues with those, you should be OK.
I did do one thing, I have an SD card inside with HBCD and the proper drivers for the system, this way I can do firmware and bios updates on stuff. I just leave the card in there now in case I need to boot into Windows for something. Takes another 5 seconds to boot, but not a big deal after the length of time these servers normally take to start up.
If you don’t already buy it, the advanced iLO license is nice to have, not sure what the cost would be but might be worth having.
Running on DL360 Gen10, without any issues.
System has P408i array card - two ssd mirrored boot drives, two 2TB data drives (mirrored)
We backup the virtual machines to a truenas server using nfs shares
Just in case anyone else comes here wonder the same thing I confirmed with Vates that the new raid cards in the G11 servers(MR416) are NOT compatible with XCP-NG… YET…
We setup a demo and in turn gained access to support.
“these new tri-mode NVMe capable RAID cards only present a 4Kn sector virtual drive, not 512, and SMAPI with XCP-ng is not currently capable of using 4Kn drives”
“trust us, 4Kn support (a whole new storage stack entirely, SMAPIv3 to be specific) is top on our priority list”
On the bright side I did find out G10’s are still in production so we can just purchase new G10’s 
You can’t replace the controller card in the gen11 for an older technology? I know the answer is “it depends” but it might be more flexible that way.
I’m sure that would work but we’d rather stick with the manufacturers hardware setup, i’m 99% sure by the time our new g10’s go out of warranty xcp will support the newer raid cards. Plus i’ve had fantastic luck with our g10 servers so i don’t mind running them a little longer
BPSTravis - FYI that MOST Gen10 models will stop being for sale in a few months, Gen10 Plus (Ice Lake) maybe slightly longer.
The Exception is the DL360 and DL380 Gen10 Plus which were extended into 2025.
If you need more specific dates on a certain model, feel free to drop me a line and I can check for you.
Hello everyone
We are using a brand new Proliant 385 Gen11 and XCP-NG runs great.
Buuut there is a “HPE MR408i-o Gen11” build in and not a 416.
Maybe this information is helpful for anybody 
I’d guess 408 vs 416 is just the internal header count of the controller, interesting that it works, have you tested full functionality yet?
No.
To be honest, i dont know how.
I did setup my four 3,84TBs SSDs with iLO to Raid 6 and done. It just works. I planing to switch from Raid 6 to 2x2 Raid 1.
Would it help, when I setup a vm and do a harddrive benchmark?
Vates support made it sound like the G11 raid controllers were completely incompatible with XCP-NG due to the block size they present to the OS so i’m not sure how you even got it to install, maybe they missed something and these controllers are compatible?
Can you see if your virtual disk reports what sector size it’s using? It must have an option to present as a 512 sector size which would be compatible.
That sure looks like a 512 sector size off your G11 MR408i, i’d say it would appear they are compatible, unless for some reason the 16 port variant doesn’t offer that same option, i can’t imagine that would be the case though. I’ll pass this along to vates support so they know some of these cards do appear to be compatible.
No matter what the native disk block size might be, the RAID config should be able to accommodate the smaller sizes, at least that’s my guess. My thoughts are that the controller can just cache eight 512 packets to make the 4k physical block
And this might only work in RAID 3, 5, 6 modes, mirror might not be able to work with the smaller block size commands. Again just a guess.
Or did HP release a different firmware that allows the smaller block sizes?
But as newer hardware comes online, it’s a change that will need to be made.
Hello everyone
I wiped everything and changed the raid to two times raid 1. Everything runs smooth and the installation was no problem at all.
Sectorsize is 512 / 4096 bytes (Logical/Physical).
My fresh installed test vm is working great.
Good to hear, thank you for testing/verifying.
