New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Does windows server on Ryzen always have issues?
Over the last few months, I tried a couple of Ryzen servers with Hetzner. Every time I keep getting occasional BSODs, or server losing connectivity, or sometimes tasks like video transcoding would stop for no reason, etc. Never had on of these issues with my Intel servers.
The only Ryzen configs I tried were AX line from Hetzner, but I wonder if anyone else happens to run into such issues sometimes with Ryzen?
Comments
I have a Ryzen laptop - running it like a champ - never seen a single BSOD on that.
Ofcourse - its not Hetzner - but, +1 for Ryzen and Windows combo
Have a AX with Hetzner aswell running a Windows only software and no problems.
Same with me, Ryzen Laptop and 2 Ryzen Desktops without problems!
Sounds like I always had bad luck with my ryzen servers
Never had a single issue with Ryzen + servers.
I have a dell laptop with Ryzen CPU and just 4 GB memory … running fine for last 6+ years
Windows 10 professional … will upgrade to windows 11 OS in next couple of months
And one more thing, I connect it to 4K external monitor
Without knowing what error you get in blue screen, it's difficult to know where your problem is.
I just had a windows 11 test PC BSOD until i confirmed brand new RAM was bad and no issues after replacement.
What are your blue screen errors?
Have you tested the RAM?
@TimboJones
I have no ryzen servers atm. They were productions servers, I didn't have the luxury to debug so I use Intel now.
Some came from network card, changing drivers, etc never solved the problem. But I had the same network card with an Intel server and same drivers and it was fine, which is why I blamed it on same HW incompatibility.
But even if the network card was the culprit for that particular servers, It is not the issue for other 2 I had.
Also, it does not explain why video transcoding operations would suddenly stop (without giving any errors). Sadly these were user reports and I didn't have time to reproduce before cancelling servers myself and moving users on Intel servers.
I know I am not giving any useful details to debug issues here, but It is not what I am looking for either, What I mainly wanted to see is if other people encountered issues with Ryzen configs too or if it was my bad luck.
@afn I believe this is specific to Hetzner Ryzen CPU + Motherboard setup. I have experienced those random crashes myself and I have seen the reports from other users as well. It is not just Windows but Linux as well.
Some nodes crash more often while the others crash less. You can try your luck and ask Hetzner to replace the node while keeping the drives.
Ok, just realize that's completely pointless and irrelevant whether other people do or do not have issues on other hardware and setups. You have a hardware or driver problem specific to your setup and absolutely no details to compare with others.
It's even worse dealing with second hand info.
Works for me at hetzner, no issues
@TimboJones well i think it does related to hardware issue. i was driven insane by it.
the only clue i manage to find it related to motherboard with unsupported DIMMs.
after 1.5 months and 6 freezes i have migrated this server to other server from PX line.
Well happened to me to, your motherboard brand is asrock rack?
No one have mentioned about Windows Activation, but do you activate your Windows?
and do you use Hetzner provided ISO or your own ISO?
It's unrelated Hetzner forum and alot of people suffer from weird freezes that require cold reboot by hetzner support with AX line within Hetzner.
I have got many BSOD with Ryzen 2500U, but it looks like linux is more stable on Ryzen laptop than Windows 10.
(Not the op)
Not a fan of ASRack. Had one H87 motherboard with PCIe issues. Went through troubleshooting with actual engineer! But couldn't reproduce the issues. One of those things where so much time and troubleshooting and in the end, swapping motherboards to another model just worked.
Random BSOD is nearly always RAM related. Other issues generally have same BSOD errors, like bad boot drivers (ie, trying to move from MBR to GPT and hardware change same time).
I get it with all sorts of hardware when its got issues
Also any details on which BSOD's the OP getting? Majority of these problems (Random reboots with a BSOD, application hang/crash, connectivity loss,etc) is likely RAM, but depending on BCC and the faulting module - you could be looking at a driver related issue.
You could actually just look at the event log (Check the 'bugcheck' source, or read the minidumps) gather useful diagnostic information from the bluescreens. You can also see if you're getting WHEA events.
No one mentioned it because its the least likely component to be causing blue screens. The best trick it does is shut down every hour til you sacrifice a virgin.
Did you save the text of the BSoD? It'd be useful to check if it's in a driver or if it's in the Windows kernel itself. A BSoD could also be due to faulty RAM (if you've got ECC RAM and it detects unrecoverable corruption, or it's non-ECC RAM that gets corrupted enough to cause issues). You really need to know which driver triggered the BSoD, and what the stop code was, to be able to debug it.
I Think I did not phrase it clearly, but I mainly mentioned using hetzner servers to ask about people using similar configs, (with Hetzner or some other provider) and mainly using windows server. Because sometimes, drivers get released for normal windows 10 but they refuse to install on windows server (happened to me once) or there might be other win server/win 10 differences that I am not aware of... Which is why I am interested in people running servers, or at least doing video editing on their PCs. So a feedback for Ryzen laptop would be irrelevant to my question.
Hmmm, where have you exactly seen these "alot of people", I am curious
--
@CheepCluck , @Daniel15, @TimboJones , and everyone else, thanks for your input guys, and to answer the question about BSODs: I remember I no longer have them sadly.
The main reason I asked is not solve an issue that no longer exists (because I no longer have these servers), but it's mainly to hear people experiences and see if I should give a Ryzen based configs another try and possibly see which specific HW commonly cause issues so I can go from there in case I run into issues. So far I see that is better to avoid ASRack MB, and check ram.
slightly off-topic rant (that could be skipped
Honestly, It is a shame that my experience with AMD was always terrible (even outside servers, for my personal laptops), because nowadays they give insane performance/price ratio.
Also, if you look at their power usage you can have 12 cores ryzen with tdp 105W but all similar intel CPUs with 10 cores or more are way more than that (even the ones with no iGPU).
Example: ryzen-9-3900x VS i9-10900kf (less performance, more TDP) and xeon W-3323
I Used several AMD processors between 2005-2007, it was complete trash compared to cheaper intel CPUs and the support I got was terrible. I swore I am not trying anything from AMD again. And starting from 2007+ I kept advising every one who turned to me for advice not to use AMD. Every morning I would wake up, eat my breakfast, curse AMD, then turn on my Intel PC. around 2018 I saw the insane value/price ratio for AMD and decided to try AX line servers. My experience was trash again. So here I am again, almost 2022 and I wonder if I should waste some more of my time on them...
I forgot to answer the original question, and haven't tried Windows Server, but Windows 10 is working well on my desktop PC which uses a Ryzen 5900x. No issues.
I guess it's a bit before that time, but I remember the AthlonXP processors being better (cheaper + faster) than the Pentium III processors of the same era. I used AMD from that era (2002-2003-ish?) until the Core 2 Duo / Nehalem Core i5 were launched (2007-2008 I think)
Hetzner does test the servers, as far as i know they didnt manage to reproduce the issue.
Also in Hetzner forum there are few topics related to AX lines with random freezes/crashes.
i think hetzner should investage this issue with the hardware, like the old I9900k had random kernel freezes and after Hetzner found the issue they got a new BIOS update that fixed it.
My hunch its related to asrack motherboard in the AX servers.
@Hetzner_OL Katie as you also replied in my previous post once i asked simillar question, and there are people suffering from same issue?
you can see there is clearly something going on with AX line, i think it worth forwarding it to the hardware team to investigate?
It's unrelated to windows as it happen to people with centos, ubuntu and etc..
I also were clueless regards it cause debug hardware issue remotely is extremly hard.
when i had the issue kernel even didnt panic, just server were frozen and we used linux.
As far as i manage to research over google, i found it related to asrack motherboard and unsupported DIMMs? but could be a bug in the motherboard specific model itself that if asrack would develop a patch to it and BIOS update be amazing solution to it.
You can see in Hetzner forums also some people reporting issue with stabillity with the AX line.
Hi there LET reader, This is my advice for any situation, regardless of the server model. If the server seems to repeatedly have trouble, I would really suggest to communicate clearly with the support teams about what is going on. Give them as much information as possible to demonstrate that you are having problems. Explain in detail what troubleshooting you have done to eliminate all other possible explanations. You can also ask them to do a full hardware check. And if you still experience issues, you can ask if they are willing to replace the server. (It doesn't hurt to ask, but I can't guarantee they will say yes.)
Naturally, if it does appear to be a hardware issue, our team will replace it; they're reasonable and will work with you. If an issue appears to be from a manufacturer, our team will need your detailed information to pass onto the manufacturer. Our team may or may not be able to confirm with you whether there is an issue with the manufacturer -- sometimes there are non-disclosure agreements that get in the way. --Katie
Thanks for your reply.
I Know Hetzner can and they do test servers if requested, but in my case Hetzner never found the issues with any of my servers with their testing method (be it Intel or AMD). I am not complaining, they end up replacing the server and solving the problem eventually as Katie said, but they don't always pin point the culprit. I am however curious to find the culprit not just replace the server.
I once had a 990x that would BSOD for no reason. They Replaced the server, kept drives, still BSOD. We concluded it is a piece of hardware with a driver that got a bad update that broke it . (or maybe the raid 0 config I created from the disk management in Windows was acting up for some reason).
One of my 8700, started acting up. it has been tested by them, they found nothing, I asked to replace ram, they agreed, still BSODs, we replaced the server in the end.
But, these are 2 or 3 cases out of 30ish (or maybe more) Intel servers I have/had with them.
As for my AX servers, I have only had 3, and all three always had crashes. I requested tests for one of them, They found nothing. Which is why we concluded it is windows drivers issue or maybe something about either the CPU or the other components they use, thus this thread.
CentOS 7 doesn't work with Ryzen 5000 series for example. kernel panics.
Depends on which ryzen series you got, 1000,2000 and 3000s should be good. 5000 might have issues.
Did you yum update? You got some other problem.
There's no more centos 7 updates regarding new hardware. You need to use CentOS8/Rocky Linux8. If you check there are many forum post regarding the issue.
@afn
I have similar issue with AX-line. When support tested server with quick test, he found nothing.
I have many nodes with AX-line. I experiens random down more than 5 servers.
I installed KVM virtualization on my servers.
Now, i have 2 AX-line servers which have similar issue.
When i tried to move my vps to another machine, i found clue that i never found random down after i moved many vps from this servers.
I tried to create vps on this 2 servers. this servers are AX-41 with 64GB RAM (non ECC).
When i created VPS with total RAM allocation are half of specification, everithing work properly.
No random down again .
This servers have uptime between 210 days until 220 days without down.
so i am sure this issue related with dimm driver.
@Hetzner_OL
Please test from your side base of my clues. i hope it could help you to find the issue.
I find some thread related with this random Shutdown on your AX-line.
my newest server on hetzner are using Intel. But i saw that you reduce EX-line currently. i don't find similar issue with EX-line. i only found issue with broken RAM and i could identify it easy. i still cannot order new AX-line if you could not find caused this issue.
I had used latest kernel 5 with centos 7, base of hetzner suggestion. I also tried used Almalinux with kernel 4 but had same issue.
Hi there, I'd kindly ask you to get into touch with our support team via [email protected] or to raise a ticket through our ticket system. Our technicians there can give you a detailed answer
-Lea
I have Ax41 , AX101 , Ax161
all running windows server 2019 from year and working flawless , all are production servers and almost all time under load.
If the BSOD indicates problems with a network interface and that interface is actually an emulated device in a VM, maybe they could expose the network via a different virtual device type instead? kvm for instance supports several types of NICs.
Can confirm never had issues with zen2 Ryzen or Eypcs on Windows 2019 or 2022 with KVMs with virtio installed, most clients use Windows on our nodes to.