Minisforum ms-01 i9-13900H 96GB with Proxmox

[lastnerdstanding]

Follow-up:

I re-purchased an MS-01 from Amazon and received it last week. I have it setup the same way except I was able to buy the bare bones version this time around. Definitely seems like a new revision though not sure if any hardware revisions were made.

I have not seen any reboots in the logs or stability issues so far. I do still see split locks on occasion. They do show up immediately on boot of my Windows VMs and one or two when they're running. I might actually attribute the split locks to the anti-virus running on them though not 100% certain.

I can't rule out any Proxmox updates doing anything either since I'm now at 8.1.4. but I feel I got a bad unit first time around whether that's bios updates or otherwise. Current microcode is a x410c which is the same as the original unit. I'll leave it on this version unless there's a need to update it.

I'll post again if I start seeing any stability issues.

 

[nicoska]

Follow-up:

...

Thanks for sharing. I'm going to re-buy again the barebone version. The first time mine came with CPU issues (random reboot under heavy load).

 

[anewsome]

Hello fellow ms-01 users. This seems like the thread with the most users experiencing Proxmox crashing issues for the ms-01. I received my order from Minisforum and promptly installed ram, nvme and Proxmox. I'm pretty familiar with Proxmox, as I've been using it for quite a while now and never seen a crash. The MS-01 are crashing pretty consistently. About 1-2 days has been the most uptime I've seen.

Here's my config:

5x MS-01 i9-13900H
96GB DDR5 5600 (each node)
3x Crucial T500 (each node, slow slot for boot, the other 2 for ceph)
2.5gb used only for AMT/vPro mgmt
2.5gb used for vmbr0
2x 10gb LACP bond1 used for Ceph

I have no pcie card installed and haven't done any special BIOS tweaks. After a few days of crashing, I did install the intel-microcode from the non-free repo, but that didn't help.

The prior microcode was 0x410c and the new microcode is 0x411c.

Proxmox version is 8.1.11, installed just two days before the 8.2 was available.

Has anyone else been able to fix the consistent crashing of Proxmox? Does disabling the E-cores fix the issue? Seems rather drastic to completely disable so many cores in the cluster.

Has anyone tried upgrading to 8.2? Do these crashes still persist? Any help you all can offer would be greatly appreciated.

Also note, this Proxmox cluster has 5x i7 Intel NUCs that have been running for more than a year without a single hiccup and no crashes. Irrelevant to my problem with the MS-01, but interesting.

 

[lastnerdstanding]

Hello fellow ms-01 users. This seems like the thread with the most users experiencing Proxmox crashing issues for the ms-01. I received my order from Minisforum and promptly installed ram, nvme and Proxmox. I'm pretty familiar with Proxmox, as I've been using it for quite a while now and never seen a crash. The MS-01 are crashing pretty consistently. About 1-2 days has been the most uptime I've seen.

Here's my config:

5x MS-01 i9-13900H
96GB DDR5 5600 (each node)
3x Crucial T500 (each node, slow slot for boot, the other 2 for ceph)
2.5gb used only for AMT/vPro mgmt
2.5gb used for vmbr0
2x 10gb LACP bond1 used for Ceph

I have no pcie card installed and haven't done any special BIOS tweaks. After a few days of crashing, I did install the intel-microcode from the non-free repo, but that didn't help.

The prior microcode was 0x410c and the new microcode is 0x411c.

Proxmox version is 8.1.11, installed just two days before the 8.2 was available.

Has anyone else been able to fix the consistent crashing of Proxmox? Does disabling the E-cores fix the issue? Seems rather drastic to completely disable so many cores in the cluster.

Has anyone tried upgrading to 8.2? Do these crashes still persist? Any help you all can offer would be greatly appreciated.

Also note, this Proxmox cluster has 5x i7 Intel NUCs that have been running for more than a year without a single hiccup and no crashes. Irrelevant to my problem with the MS-01, but interesting.

This sounds very similar to my issues on the first first batch. It would randomly reboot (logs indicated a reboots when I didn't initiate them) so it told me it was unstable. I switched back to my old NUC which was originally running ESXi before converting it and experienced no stability issues. I opted to return my unit and wait for a new batch. I also updated the microcode with no luck to stability (0x411c) No PCI devices and was using the same hardware minus the included M.2. I upgraded to 64GB. I'm only running Windows & Win Server on 4 VMs. No heavy loads but I also wasn't on heavy loads prior.

I haven't updated to 8.2 yet but I would try it and see if there's any differences. I don't recall which version I was on previously. Maybe 8.1.2? So my jump from 1.2 to 1.4 probably wouldn't make a difference. I still feel it's a hardware issue somewhere or possibly a BIOS issue. So far it's been ok with my replacement unit.

I can't say disabling e-cores would help or not but I would try it. If you're still in a return window from wherever you got them, might be worth a try to swap out with the new batch. There's box art now when before it was just a plain black box.

 

[SBMe]

Yeah I'd be very interested if anyone has upgraded to the new Proxmox 8.2.1 that was just released on April 24, 2024 along with their backup server to 3.2 on April 25th ( same page ) ?

I was just reading last night that the MS-01 13th Gen has more problems on Level1Techs and knew this was also a great place to know what's going on.

I watched the new official YT Proxmox video last night, they did it way too fast though, but a lot of nice improvements.

I just noticed late last night that Amazon has the MS-01 back in stock again for the 12th & 13th Gen barebones and for the ones with 32GB of RAM.. with big rebates.
Should NOT have to disable E-Cores to get this thing to work, I want to use mine for a bunch of containers so that's not going to work. I was waiting for Amazon for the return possibility, just in case and also to get an extra 5% off with their card, every bit helps. If need to disable things then I'll just do a build myself I think? I should look to see if anyone has purchased numerous i5 versions, looks like they are always out of stock, I assume they only sell like 10-20% of those compared to the other 2 models but who knows what if any business orders they are getting.

 

[JaxJiang]

Hello fellow ms-01 users. This seems like the thread with the most users experiencing Proxmox crashing issues for the ms-01. I received my order from Minisforum and promptly installed ram, nvme and Proxmox. I'm pretty familiar with Proxmox, as I've been using it for quite a while now and never seen a crash. The MS-01 are crashing pretty consistently. About 1-2 days has been the most uptime I've seen.

Here's my config:

5x MS-01 i9-13900H
96GB DDR5 5600 (each node)
3x Crucial T500 (each node, slow slot for boot, the other 2 for ceph)
2.5gb used only for AMT/vPro mgmt
2.5gb used for vmbr0
2x 10gb LACP bond1 used for Ceph

I have no pcie card installed and haven't done any special BIOS tweaks. After a few days of crashing, I did install the intel-microcode from the non-free repo, but that didn't help.

The prior microcode was 0x410c and the new microcode is 0x411c.

Proxmox version is 8.1.11, installed just two days before the 8.2 was available.

Has anyone else been able to fix the consistent crashing of Proxmox? Does disabling the E-cores fix the issue? Seems rather drastic to completely disable so many cores in the cluster.

Has anyone tried upgrading to 8.2? Do these crashes still persist? Any help you all can offer would be greatly appreciated.

Also note, this Proxmox cluster has 5x i7 Intel NUCs that have been running for more than a year without a single hiccup and no crashes. Irrelevant to my problem with the MS-01, but interesting.

Hi. We are watching for crash problem. If you encounter this problem, official support is available for return and exchange.
We officially take this issue very seriously.
Becasue of shipping time delay, It would be a little later for the technical analysis department to receive and analyze the problematic machines.
We just got a machine that is said to have a crash problem.
Our official staff are doing their best to analyze this problem and do their best to come up with a way to solve this problem without transportation.

 

[anewsome]

Thanks @JaxJiang . Trust me, I'm watching closely too. Since I've already migrated all my VMs to the new cluster with the MS-01, all my critical services are running on them. Crashes and frozen Proxmox hosts are a real pain. Since my post yesterday, I haven't seen a crash. That's more than 24 hours without a crash. One of the 5 has been up for a whole 3 days. Here's the uptime report:

ms01 19:08:16 up 1 day,  4:49,  0 user,  load average: 0.74, 0.69, 0.66
ms02 19:08:17 up 1 day, 20:47,  0 user,  load average: 0.29, 0.33, 0.29
ms03 19:08:17 up 3 days, 38 min,  0 user,  load average: 0.28, 0.33, 0.37
ms04 19:08:17 up 1 day, 11:11,  0 user,  load average: 0.77, 0.55, 0.47
ms05 19:08:18 up 1 day, 21:39,  1 user,  load average: 0.46, 0.39, 0.45

 

[JaxJiang]

Thanks @JaxJiang . Trust me, I'm watching closely too. Since I've already migrated all my VMs to the new cluster with the MS-01, all my critical services are running on them. Crashes and frozen Proxmox hosts are a real pain. Since my post yesterday, I haven't seen a crash. That's more than 24 hours without a crash. One of the 5 has been up for a whole 3 days. Here's the uptime report:

ms01 19:08:16 up 1 day,  4:49,  0 user,  load average: 0.74, 0.69, 0.66
ms02 19:08:17 up 1 day, 20:47,  0 user,  load average: 0.29, 0.33, 0.29
ms03 19:08:17 up 3 days, 38 min,  0 user,  load average: 0.28, 0.33, 0.37
ms04 19:08:17 up 1 day, 11:11,  0 user,  load average: 0.77, 0.55, 0.47
ms05 19:08:18 up 1 day, 21:39,  1 user,  load average: 0.46, 0.39, 0.45

Hi. We will release a new BIOS to try fix this problem.
Recently, our maintenance department and a remote customer fixed this problem by reflashing the BIOS.

 

[binboum]

Hi. We will release a new BIOS to try fix this problem.
Recently, our maintenance department and a remote customer fixed this problem by reflashing the BIOS.

I notice the same issue for me, with one of the batches received in February.

When the BIOS update is released, could you inform us about it here?

 

[JaxJiang]

I notice the same issue for me, with one of the batches received in February.

When the BIOS update is released, could you inform us about it here?

Of course. I will notify everywhere where i can see it.

 

[makyz]

I am having the same issues. I have not been present but when I come back to the machine, its off and not responsive until I power it down and restart it. Are there some diagnostic commands I can run that could help with the debugging?

Notes:
- started happening after I installed proxmox 8.2.2 and added a couple of kubernetes nodes with light home/media automation containers
- Intel Core i9-13900H, 1x nvme (https://www.amazon.com/dp/B09QV5KJHV), 2x 48GB RAM (https://www.amazon.com/dp/B0C79T2CP7)
- I have applied the microcode update (Firmware Updates - Proxmox VE), but it still locks up. Screenshot below of current version:



I turned off the efficiency cores (see resulting proxmox screenshot) this morning to see if the machine locks up again, I will report back:

 

[JaxJiang]

@makyz @binboum Hi You can download BIOS on minisforum
This is not the lastest BETA BIOS. But may fix your problem.
If it works or you still have problems, please give us feedback.

 

[makyz]

@makyz @binboum Hi You can download BIOS on minisforum
This is not the lastest BETA BIOS. But may fix your problem.
If it works or you still have problems, please give us feedback.

I already have V1.22 (see photo below), is this just a different BIOS with the same version number?

 

[SBMe]

Well that's weird, because it says it just posted as 1.22 Final as today 050724 or 05-07-24.

Wait, can you only update it through Windows? I checked the other options and they went blank.

BIOS V1.22

Windows

05/07/2024

13.85MB

 

[JaxJiang]

@makyz
If you have hang or freeze problem. Just flash it.
This is not lastest internel version of BIOS.

@SBMe
You can download it. It can upgrade in UEFI or windows.

 

[makyz]

Follow up from: https://forums.servethehome.com/ind...i9-13900h-96gb-with-proxmox.43380/post-425708

Experienced another lock up overnight with all efficiency cores disabled. This time I left it connected to a monitor so I got some errors:




After some googling, I am going to try disabling all c states and re-enabling all my Active Efficient-cores, and see if it locks up overnight again.

If I still lock up, I will disable all effecient-cores again and test again.

and if THAT doesnt work, I will flash the BIOS

 

[makyz]

Follow up from: https://forums.servethehome.com/ind...i9-13900h-96gb-with-proxmox.43380/post-425708

Experienced another lock up overnight with all efficiency cores disabled. This time I left it connected to a monitor so I got some errors:

View attachment 36528
View attachment 36529

After some googling, I am going to try disabling all c states and re-enabling all my Active Efficient-cores, and see if it locks up overnight again.

If I still lock up, I will disable all effecient-cores again and test again.

and if THAT doesnt work, I will flash the BIOS

Two days of uptime without freezing up, I believe disabling C states has solved my problems.

I will start the research on keeping C states enabled since there are power draw benefits. There are already some threads about C state compatibility with proxmox.

 

[anewsome]

Two days of uptime without freezing up, I believe disabling C states has solved my problems.

I will start the research on keeping C states enabled since there are power draw benefits. There are already some threads about C state compatibility with proxmox.

I had over 10 days of uptime across 5 MS01 proxmox nodes and then one crashed today. Since the are headless with no monitor, it's impossible to see what was on the display output before the crash. Plugging in a monitor and keyboard, it's just a black screen with no response to keyboard input. vPro remote management is dead too. If I see another crash, I'll also try disabling the C-states on all the nodes.

Please do report back if you see any crashes though.

 

[makyz]

I had over 10 days of uptime across 5 MS01 proxmox nodes and then one crashed today. Since the are headless with no monitor, it's impossible to see what was on the display output before the crash. Plugging in a monitor and keyboard, it's just a black screen with no response to keyboard input. vPro remote management is dead too. If I see another crash, I'll also try disabling the C-states on all the nodes.

Please do report back if you see any crashes though.

Unfortunately, I experienced another crash this morning. Did not have it hooked up to a monitor so I don't have any errors but I will start leaving it on a small monitor to track down further issues. As you can see its only running a few k8s pods, so there isn't much load on the system. I won't be able to move some more critical workloads on it until I figure out the lock ups. I might return the unit before its too late.

This was my longest run tho (refer to graph above).

ready for the next crash!

 

[PigLover]

View attachment 36608

I can't help with the crash (sorry) but that is a cute little monitor! If I didn't know how small the MS-01 is you'd almost miss how tiny that monitor must be