MinisForum MS-01 : heating problem

[BlueChris]

One issue with cooling is also dust accumulation, leaving the MS-01 with any kind of load will mean you have to clean it of dust regularly.. That's an issue for a 24/7 server..

yes you are right on this. After 15 days of running, the fan of the CPU has dust and the noise is raised. As a matter of fact i cleaned it yesterday and now i will put in the Sleeve Case of MS-01 a mesh or something inside to the top and bottom of it to hold the dust.

I am thinking to take out the sleeve totally and put it inside an HTPC case... after all i have an eGPU to m.2 card on it and i have an Nvidia 1070ti on the machine since yesterday... if i do what i say i will end up with a single box with the MS-01 on it and everything else also with the addition of some normal fans and filters....
Still trying to decide what box to buy.. i have a Silverstone GD10 laying around but i think is too big for what i need.. even though i have it with a LAMPTRON CW611 which is great because it has fan control with temperature etc...

 

[heromode]

I do stand by my opinion that the cooling is flawed. If you keep this machine running on a desk in a room where people also reside, the MS-01 will basically grind to a halt in about 6 months or something.

You can't just go from a standard ATX case to a minicase like this by just cramming all components in there, and also just use mini sized fans, because the dust particles and dead skin cells stay the same.

Only after i got one of those dyson type stick vacuum cleaners with a clear plastic dust container did i ever realize the insane amount of dust a small apartment generates in just a month.

The only viable solution i can think of is external fans, probably in a push-pull configuration, with internal airducts running straight through the case, and designed in such a manner, that when the air flows through the box inside the ducts, there are small vent holes in the sides of the ducts that generate a negative air pressure in the space surrounding the ducts as air flows by, where the hot components are, so that the hot air around the components gets sucked into the ducts, and exits.

Because you can't just have a external case fan blowing straight into the case, and onto the components either, that would just fill the case with dust, and cover the components with dust in notime.

 

[Phenic]

I do stand by my opinion that the cooling is flawed. If you keep this machine running on a desk in a room where people also reside, the MS-01 will basically grind to a halt in about 6 months or something.

You can't just go from a standard ATX case to a minicase like this by just cramming all components in there, and also just use mini sized fans, because the dust particles and dead skin cells stay the same.

Only after i got one of those dyson type stick vacuum cleaners with a clear plastic dust container did i ever realize the insane amount of dust a small apartment generates in just a month.

The only viable solution i can think of is external fans, probably in a push-pull configuration, with internal airducts running straight through the case, and designed in such a manner, that when the air flows through the box inside the ducts, there are small vent holes in the sides of the ducts that generate a negative air pressure in the space surrounding the ducts as air flows by, where the hot components are, so that the hot air around the components gets sucked into the ducts, and exits.

Because you can't just have a external case fan blowing straight into the case, and onto the components either, that would just fill the case with dust, and cover the components with dust in notime.

I'll have to disagree. Cooling could be better but isn't flawed. There are many tiny/micro/mini homelabbers here at STH and we run basically the same form factor as the ms-01 and our machines are running fine for years.

 

[BlueChris]

I do stand by my opinion that the cooling is flawed. If you keep this machine running on a desk in a room where people also reside, the MS-01 will basically grind to a halt in about 6 months or something.

You can't just go from a standard ATX case to a minicase like this by just cramming all components in there, and also just use mini sized fans, because the dust particles and dead skin cells stay the same.

Only after i got one of those dyson type stick vacuum cleaners with a clear plastic dust container did i ever realize the insane amount of dust a small apartment generates in just a month.

The only viable solution i can think of is external fans, probably in a push-pull configuration, with internal airducts running straight through the case, and designed in such a manner, that when the air flows through the box inside the ducts, there are small vent holes in the sides of the ducts that generate a negative air pressure in the space surrounding the ducts as air flows by, where the hot components are, so that the hot air around the components gets sucked into the ducts, and exits.

Because you can't just have a external case fan blowing straight into the case, and onto the components either, that would just fill the case with dust, and cover the components with dust in notime.

I'll have to disagree. Cooling could be better but isn't flawed. There are many tiny/micro/mini homelabbers here at STH and we run basically the same form factor as the ms-01 and our machines are running fine for years.

This... the point is with what you compare the MS-01. If you compare it with what we have in our server rooms in work then yes the temp is an issue if you have a whole system with full nvme's to sit at 55-60c where the server room has everything at 20-22, but if you compare it with other tiny pc's is perfectly fine.

Either way.. for the disks i took out the 120mm that i had on the bottom of it and i just puted there a laptop stand that has 1 big fan... same result but more neat.

 

[heromode]

This... the point is with what you compare the MS-01. If you compare it with what we have in our server rooms in work then yes the temp is an issue if you have a whole system with full nvme's to sit at 55-60c where the server room has everything at 20-22, but if you compare it with other tiny pc's is perfectly fine.

Either way.. for the disks i took out the 120mm that i had on the bottom of it and i just puted there a laptop stand that has 1 big fan... same result but more neat.

but the MS-01 is packed with high performance components, CPU, GPU, M.2 SSD, 2x 10 GB Ethernet, designed to run 24/7 at temps between 70-90 degrees, pulling hundreds of watts. If you do that, then it will eventually grind to a halt, because it will be filled with dust, and the cooling system will fail.

That's a flawed design.

If you buy a compact Porsche with a high performance engine, you will actually be able to drive it at high speeds for as long as you have gas. If you could only drive it fast for 5 minutes, and the rest of the way you could only drive it like a Toyota Prius, then you would be pissed off, and return the car as defective.

 

[ms264556]

but the MS-01 is packed with high performance components, CPU, GPU, M.2 SSD, 2x 10 GB Ethernet, designed to run 24/7 at temps between 70-90 degrees, pulling hundreds of watts. If you do that, then it will eventually grind to a halt, because it will be filled with dust, and the cooling system will fail.

That's a flawed design.

There's no GPU unless you add one & the CPU is a 45W unit. Nobody runs these mini PCs at 100% all the time. my MS-01 is currently running pfSense & Ruckus SmartZone on proxmox into a 10G switch with a DAC. I'm currently streaming a video using the iGPU for transcoding & there's ~400Mb continuous traffic from APs going into the SmartZone then into pfSense. The internal temperature is 38° and I can't hear the fan, it's running so slow.

When I did have a GPU in it, as my main PC, the good separation and independent blower-style airflows on each side meant that both the GPU and the CPU were cool and had their fans running at minimal speed.

I've picked up a bunch of 1L mini PCs which have usually been running for 3-4 years without any cleaning. The fans are usually a little furry & they can sometimes have noisy bearings, but have never been full of dust or ground to a halt.

 

[heromode]

There's no GPU unless you add one & the CPU is a 45W unit. Nobody runs these mini PCs at 100% all the time. my MS-01 is currently running pfSense & Ruckus SmartZone on proxmox into a 10G switch with a DAC. I'm currently streaming a video using the iGPU for transcoding & there's ~400Mb continuous traffic from APs going into the SmartZone then into pfSense. The internal temperature is 38° and I can't hear the fan, it's running so slow.

When I did have a GPU in it, as my main PC, the good separation and independent blower-style airflows on each side meant that both the GPU and the CPU were cool and had their fans running at minimal speed.

I've picked up a bunch of 1L mini PCs which have usually been running for 3-4 years without any cleaning. The fans are usually a little furry & they can sometimes have noisy bearings, but have never been full of dust or ground to a halt.

FINE. I admit that the MS-01 is a nice little box. I really do.

But maybe someone here that has alot of these would be willing to put one to the test. 24/7 Run stress-ng on the CPU, hammer all nic interfaces with iperf3 continuously. use some disk benchmark to hammer all nvme SSD's. Transcode on the igpu. I wanna know what happens after a few months

 

[BlueChris]

FINE. I admit that the MS-01 is a nice little box. I really do.

But maybe someone here that has alot of these would be willing to put one to the test. 24/7 Run stress-ng on the CPU, hammer all nic interfaces with iperf3 continuously. use some disk benchmark to hammer all nvme SSD's. Transcode on the igpu. I wanna know what happens after a few months

m8 the MS-01 is like a laptop in a USFF form. Its not designed to do 24/7 burn in tests. Who in his right of mind bought MS-01's to melt them in a production environment? i think somehow you have it wrong in your mind.. its a tiny tiny full feature pc that its perfect for AIO Lab needs in our homes but cannot stand for certain constant high loads.
What we do some people with extra fans is that our OCD kicking and we try to lower a bit the overall temp of the box to be fine.
Personally? its my 1st super chinese machine that i run and i am sceptical if it holds ok in time.. but not from the heat, i have more fears of the quality of the materials that is build if they fail in some years. So far i had servers in home or high quality motherboards for my lab.

 

[heromode]

I admit to trying to find something negative about the MS-01 for no good reason.. Personally i'm still waiting for a western company to copy the idea, and bring a high quality version of the same concept to market.

It's funny because it used to be the other way around.. The west would innovate, and the chinese would produce a low quality knockoff
Now the chinese have innovated the MS-01, and i'm waiting for a higher quality western knockoff

 

[BlueChris]

I admit to trying to find something negative about the MS-01 for no good reason.. Personally i'm still waiting for a western company to copy the idea, and bring a high quality version of the same concept to market.

It's funny because it used to be the other way around.. The west would innovate, and the chinese would produce a low quality knockoff
Now the chinese have innovated the MS-01, and i'm waiting for a higher quality western knockoff

As long as they are late to the game the worse it is...

Erying Unveils B760M Motherboards With Up To Core i9-13900H 95W CPUs, Vapor-Chamber Cooler

Erying has introduced its brand new B760M motherboards which come equipped with up to Intel Core i9-13900H CPUs & Vapor Chamber cooling.

wccftech.com

 

[martel80]

I just found this forum while googling about the MS-01.

I do audio engineering and travel a lot. Almost all places where I go do have at least a screen with an HDMI input.
I was reading through this thread and saw ( very few I admit) some individuals being bothered or should I say concerned about heat dissipation.

My mixing sessions usually last up to 2 - 3 hours (Maximum) and I often ran out of CPU resources to mix music which mean that it usually max out my cpu (therefor, the fan and CPU run high).

Now , given those heat dissipation concerns, would it be wise for me to get one of those or should I still venture in the laptop territory just to make sure I don't get any BSOD ?

Also, how's the noise floor of the fan at low usage ?

Is it quiet enough to record vocals in the same room or its a noisy little machine ?

Thanks for the input.

EDIT:
I've just read this :

'' Idle power consumption was higher than the average mini PC. We saw around 25-29W at idle on our test system. Noise hit 37-38dba on our new 34.5dba noise floor set. The system under load would spin up to around 115W for ~45 seconds before pushing back down to 90-95W and run at that level consistently. Noise would get into the 44dba range and a bit more if it is run at 100%. ''

Would you say this is accurate ?

https://www.servethehome.com/minisforum-ms-01-review-the-10gbe-with-pcie-slot-mini-pc-intel/5/

 

[skipper ohms]

Is it quiet enough to record vocals in the same room or its a noisy little machine ?

No, the SSD fan has a high pitch which your mic will likely pick up. Plus no one will take you seriously without a MBP lol

 

[martel80]

No, the SSD fan has a high pitch which your mic will likely pick up. Plus no one will take you seriously without a MBP lol

So I understand by your stupid answer that it is a quiet machine or are you just another troll that has nothing else to do of his life then lose his time on forums so he get a bit of attention but inherently never used or owned the MS-01 ? Plus you can't be taken seriously because you said MBP lol

Obviously, if an intelligent human being on this forum has hands on experience, I'd be a lot more respectful of his answer and very appreciative.

 

[ms264556]

So I understand by your stupid answer that it is a quiet machine or are you just another troll that has nothing else to do of his life then lose his time on forums so he get a bit of attention but inherently never used or owned the MS-01 ? Plus you can't be taken seriously because you said MBP lol

Obviously, if an intelligent human being on this forum has hands on experience, I'd be a lot more respectful of his answer and very appreciative.

Please don't join a forum and immediately post unpleasant comments. The users here are very helpful and the mood is quite light and informal.

There was a good answer to your question: the MS-01 fans do have a high pitched whine.

(And the MBP comment made me crack a smile, having shared space with musicians for several years).

 

[martel80]

Please don't join a forum and immediately post unpleasant comments. The users here are very helpful and the mood is quite light and informal.

There was a good answer to your question: the MS-01 fans do have a high pitched whine.

(And the MBP comment made me crack a smile, having shared space with musicians for several years).

I simply responded to someone that acted like a cunt with the same tone.

Please don't welcome new members in your forum and immediately act like a cunt. I'm a very friendly user and my mood is quite light and informal.

Thank you very much for confirming that the MS-01 fan has a high pitch noise. A simple answer to a simple question. Highly appreciated.

 

[piranha32]

Please don't welcome new members in your forum and immediately act like a cunt.

You are off to a great start on the forum.

 

[Patrick]

OK enough of this tone. Please keep things civil.

 

[Joker]

Hi, perhaps somebody can help me out, I just got my MS-01 yesterday but sadly one of the fan is borked and doesn't spin well and does bad noise, I am not so fond to ship it back for ask to replace the borked fan since it takes a while and I need the box running, hence I wanted ask if you guys know and can provide me a compatible list of fans, so that I just can buy one and replace it by myself without the need to ship it back

 

[amgems]

Had my ms-01 for about a month. Running ProxMox on it, mainly opnsense. Two Samsung SSD 990 PRO 2TB, zfs mirrored. First thing I had to do was turn off Turbo mode in the BIOS. Otherwise, the CPU temps were climbing too high: up above 70C. (at some point I will try to replace thermal paste).

I periodically run `sensors` to keep track of the running temps, and this has been a good solution.

Yesterday, I was wondering why I never got any email from my PoxMox, found where it was being delivered and classified as spam.
One of particular interest was the sole indication that one of the mirror devices in the zfs root pool went bad, about 18 days ago.
I had rebooted the ms01 many times, and never noticed. One nvme device had disappeared entirely.
I migrated all my VMs off, powered down, removed one SSD, booted, and came up on the system from May 6.
`smartctl` indicated zero issues, zero errors. (need to remember the SSD-specific diag utils).
Put back the other SSD, rebooted, and both /dev/nvme were there, and `zpool` had already finished re-silvering 800GB by the time I looked.

While I was doing this, I had the case off and a portable fan directed at the SSD fan. Temperatures were OK.

Here is an excerpt of the log during reboot(?) when I had two /dev/nvme:

May 06 21:28:07 pve sensors[1228]: nvme-pci-5800
May 06 21:28:07 pve sensors[1228]: Adapter: PCI adapter
May 06 21:28:07 pve sensors[1228]: Composite:    +47.9°C  (low  = -273.1°C, high = +81.8°C)
May 06 21:28:07 pve sensors[1228]:                        (crit = +84.8°C)
May 06 21:28:07 pve sensors[1228]: Sensor 1:     +47.9°C  (low  = -273.1°C, high = +65261.8°C)
May 06 21:28:07 pve sensors[1228]: Sensor 2:     +62.9°C  (low  = -273.1°C, high = +65261.8°C)
May 06 21:28:07 pve sensors[1228]: acpitz-acpi-0
May 06 21:28:07 pve sensors[1228]: Adapter: ACPI interface
May 06 21:28:07 pve sensors[1228]: temp1:        +27.8°C
May 06 21:28:07 pve sensors[1228]: coretemp-isa-0000
May 06 21:28:07 pve sensors[1228]: Adapter: ISA adapter
May 06 21:28:07 pve sensors[1228]: Package id 0:  +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 0:        +43.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 4:        +44.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 8:        +40.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 12:       +43.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 16:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 20:       +43.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 24:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 25:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 26:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 27:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 28:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 29:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 30:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: Core 31:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 06 21:28:07 pve sensors[1228]: nvme-pci-5900
May 06 21:28:07 pve sensors[1228]: Adapter: PCI adapter
May 06 21:28:07 pve sensors[1228]: Composite:    +49.9°C  (low  = -273.1°C, high = +81.8°C)
May 06 21:28:07 pve sensors[1228]:                        (crit = +84.8°C)
May 06 21:28:07 pve sensors[1228]: Sensor 1:     +49.9°C  (low  = -273.1°C, high = +65261.8°C)
May 06 21:28:07 pve sensors[1228]: Sensor 2:     +59.9°C  (low  = -273.1°C, high = +65261.8°C)

Here is the error:

May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 464 (b1d0) opcode 0x2 (I/O Cmd) QID 4 timeout, aborting req_op:READ(0) size:12288
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 111 (806f) opcode 0x1 (I/O Cmd) QID 7 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 265 (3109) opcode 0x1 (I/O Cmd) QID 1 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 464 (c1d0) opcode 0x2 (I/O Cmd) QID 4 timeout, aborting req_op:READ(0) size:12288
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 111 (906f) opcode 0x1 (I/O Cmd) QID 7 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 265 (4109) opcode 0x1 (I/O Cmd) QID 1 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 464 (d1d0) opcode 0x2 (I/O Cmd) QID 4 timeout, aborting req_op:READ(0) size:12288
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 111 (a06f) opcode 0x1 (I/O Cmd) QID 7 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve systemd[1]: apt-daily.service: Deactivated successfully.
May 09 19:27:08 pve systemd[1]: Finished apt-daily.service - Daily apt download activities.
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 265 (5109) opcode 0x1 (I/O Cmd) QID 1 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 464 (e1d0) opcode 0x2 (I/O Cmd) QID 4 timeout, aborting req_op:READ(0) size:12288
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 111 (b06f) opcode 0x1 (I/O Cmd) QID 7 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 265 (6109) opcode 0x1 (I/O Cmd) QID 1 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 464 (f1d0) opcode 0x2 (I/O Cmd) QID 4 timeout, aborting req_op:READ(0) size:12288
May 09 19:27:08 pve kernel: nvme nvme0: I/O tag 111 (c06f) opcode 0x1 (I/O Cmd) QID 7 timeout, aborting req_op:WRITE(1) size:4096
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: nvme nvme0: Abort status: 0x0
May 09 19:27:08 pve kernel: INFO: task z_wr_iss_h:475 blocked for more than 122 seconds.
May 09 19:27:08 pve kernel:       Tainted: P           O       6.8.4-2-pve #1
May 09 19:27:08 pve kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May 09 19:27:08 pve kernel: task:z_wr_iss_h      state:D stack:0     pid:475   tgid:475   ppid:2      flags:0x00004000

On May 08 I still had two `/dev/nvme`:

May 08 02:56:49 pve kernel:  nvme1n1: p1 p2 p3
May 08 02:56:49 pve kernel:  nvme0n1: p1 p2 p3
...
May 08 02:56:51 pve systemd[1]: Started dbus.service - D-Bus System Message Bus.
May 08 02:56:51 pve sensors[1239]: nvme-pci-5800
May 08 02:56:51 pve sensors[1239]: Adapter: PCI adapter
May 08 02:56:51 pve sensors[1239]: Composite:    +49.9°C  (low  = -273.1°C, high = +81.8°C)
May 08 02:56:51 pve sensors[1239]:                        (crit = +84.8°C)
May 08 02:56:51 pve sensors[1239]: Sensor 1:     +49.9°C  (low  = -273.1°C, high = +65261.8°C)
May 08 02:56:51 pve sensors[1239]: Sensor 2:     +63.9°C  (low  = -273.1°C, high = +65261.8°C)
May 08 02:56:51 pve sensors[1239]: acpitz-acpi-0
May 08 02:56:51 pve sensors[1239]: Adapter: ACPI interface
May 08 02:56:51 pve sensors[1239]: temp1:        +27.8°C
May 08 02:56:51 pve sensors[1239]: coretemp-isa-0000
May 08 02:56:51 pve sensors[1239]: Adapter: ISA adapter
May 08 02:56:51 pve sensors[1239]: Package id 0:  +50.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 0:        +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 4:        +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 8:        +43.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 12:       +44.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 16:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 20:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 24:       +50.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 25:       +49.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 26:       +49.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 27:       +49.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 28:       +47.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 29:       +47.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 30:       +47.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: Core 31:       +47.0°C  (high = +100.0°C, crit = +100.0°C)
May 08 02:56:51 pve sensors[1239]: nvme-pci-5900
May 08 02:56:51 pve sensors[1239]: Adapter: PCI adapter
May 08 02:56:51 pve sensors[1239]: Composite:    +51.9°C  (low  = -273.1°C, high = +81.8°C)
May 08 02:56:51 pve sensors[1239]:                        (crit = +84.8°C)
May 08 02:56:51 pve sensors[1239]: Sensor 1:     +51.9°C  (low  = -273.1°C, high = +65261.8°C)
May 08 02:56:51 pve sensors[1239]: Sensor 2:     +61.9°C  (low  = -273.1°C, high = +65261.8°C)
May 08 02:56:51 pve smartd[1222]: smartd 7.3 2022-02-28 r5338 [x86_64-linux-6.8.4-2-pve] (local build)

but on May 14, there was only one:

May 14 09:36:51 pve kernel: nvme 0000:59:00.0: platform quirk: setting simple suspend
May 14 09:36:51 pve kernel: nvme 0000:58:00.0: platform quirk: setting simple suspend
May 14 09:36:51 pve kernel: i40e 0000:02:00.0: fw 9.120.73026 api 1.15 nvm 9.20 0x8000d8c5 0.0.0 [8086:1572] [8086:0000]
May 14 09:36:51 pve kernel: nvme nvme0: pci function 0000:59:00.0
May 14 09:36:51 pve kernel: nvme 0000:59:00.0: enabling device (0000 -> 0002)
May 14 09:36:51 pve kernel: nvme nvme1: pci function 0000:58:00.0
May 14 09:36:51 pve kernel: usb usb2: New USB device found, idVendor=1d6b, idProduct=0003, bcdDevice= 6.08
May 14 09:36:51 pve kernel: usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1
May 14 09:36:51 pve kernel: usb usb2: Product: xHCI Host Controller
May 14 09:36:51 pve kernel: usb usb2: Manufacturer: Linux 6.8.4-2-pve xhci-hcd
May 14 09:36:51 pve kernel: usb usb2: SerialNumber: 0000:00:14.0
May 14 09:36:51 pve kernel: hub 2-0:1.0: USB hub found
May 14 09:36:51 pve kernel: hub 2-0:1.0: 4 ports detected
May 14 09:36:51 pve kernel: nvme nvme1: Shutdown timeout set to 10 seconds
May 14 09:36:51 pve kernel: nvme nvme1: 16/0/0 default/read/poll queues
May 14 09:36:51 pve kernel:  nvme1n1: p1 p2 p3
...
May 14 09:36:53 pve smartd[1258]: Configuration file /etc/smartd.conf was parsed, found DEVICESCAN, scanning devices
May 14 09:36:53 pve systemd[1]: Started dbus.service - D-Bus System Message Bus.
May 14 09:36:53 pve sensors[1273]: coretemp-isa-0000
May 14 09:36:53 pve sensors[1273]: Adapter: ISA adapter
May 14 09:36:53 pve sensors[1273]: Package id 0:  +49.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 0:        +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 4:        +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 8:        +44.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 12:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 16:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 20:       +46.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 24:       +48.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 25:       +48.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 26:       +48.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 27:       +48.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 28:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 29:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 30:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: Core 31:       +45.0°C  (high = +100.0°C, crit = +100.0°C)
May 14 09:36:53 pve sensors[1273]: acpitz-acpi-0
May 14 09:36:53 pve sensors[1273]: Adapter: ACPI interface
May 14 09:36:53 pve sensors[1273]: temp1:        +27.8°C
May 14 09:36:53 pve sensors[1273]: nvme-pci-5800
May 14 09:36:53 pve sensors[1273]: Adapter: PCI adapter
May 14 09:36:53 pve sensors[1273]: Composite:    +56.9°C  (low  = -273.1°C, high = +81.8°C)
May 14 09:36:53 pve sensors[1273]:                        (crit = +84.8°C)
May 14 09:36:53 pve sensors[1273]: Sensor 1:     +56.9°C  (low  = -273.1°C, high = +65261.8°C)
May 14 09:36:53 pve sensors[1273]: Sensor 2:     +68.8°C  (low  = -273.1°C, high = +65261.8°C)
May 14 09:36:53 pve smartd[1258]: Device: /dev/nvme1, opened
May 14 09:36:53 pve systemd[1]: Started ksmtuned.service - Kernel Samepage Merging (KSM) Tuning Daemon.
May 14 09:36:53 pve lxcfs[1274]: Starting LXCFS at /usr/bin/lxcfs

I do not know if my issues were due to thermals being exceeded. `smartctl` shows nothing much of use.

Nevertheless, if anyone comes up with a 1/2 decent cooling option for these, other than my current "standing case on side and blowing air on it" approach, please let me know.

 

[peterhjalmarsson]

[...]
caplam the stock paste in MS-01 is really really bad people replacing it with better paste is seeing 10C which is crazy usually its like 5C. If you are going to use your box for transcoding I would go LM.

Get you some Flitz polish apply cpu/heatsink and clean it off. Apply a little LM to a piece of paper and stick the applicator in it and apply to cpu. Any excess from cpu die put on raised heatsink. If you get any LM anywhere qtip with some alcohol and it comes right off. When you use qtip to clean LM off toss qtip you don't want to clean anything with it after.

It is really easy just take your time and massage LM till it starts sticking (Flitz makes it a lot better to apply). Nothing can beat LM conductivity.

Also the stock fan curves are low in my opinion I dropped 4C by making them a little more aggressive without added noise.

I must admit to being more than a little afraid of what liquid metal could do. What do you think of the thermal pads Linus is selling? PTM7950 Phase Change Thermal Pad – LTTStore