Hi guys,

 

I recently upgraded the mobo and CPU on my main unraid server and consolidated computers to run plex on it.

I am running this in a mid-tower case with 3 5-in-4 cages. I am using the stock HSF that came with the processor which I thought would be enough to run plex, a few dockers...

 

I have noticed this in the log. Is it really overheating? Do I need to upgrade my HSF? Not sure if this is something unraid is doing in error or the CPU temp is really getting too hot and is being throttled? I've never had this happen before with previous AMD builds but never really ran plex and dockers, etc... Seems like there have been a lot of events that this has occurred?

 

Mobo: ASRock H370M Pro4 LGA 1151 

 

Apr 21 04:10:59 Sabertooth kernel: CPU1: Core temperature above threshold, cpu clock throttled (total events = 22831087)
Apr 21 04:10:59 Sabertooth kernel: CPU3: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU4: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU5: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU0: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU2: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU1: Package temperature above threshold, cpu clock throttled (total events = 30107208)
Apr 21 04:10:59 Sabertooth kernel: CPU1: Core temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU4: Package temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU3: Package temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU0: Package temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU5: Package temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU1: Package temperature/speed normal
Apr 21 04:10:59 Sabertooth kernel: CPU2: Package temperature/speed normal

 

 

Just now, maxse said:

the CPU temp is really getting too hot and is being throttled?

^^ this

Side note though

 

Stock coolers ARE more than capable of cooling the CPU if the case is also up to snuff.  I've been running the same CPU 100% maxed out for the last 3 days on a stock HSF and temps have not gotten above 63C.  You might want to check your fan profile in the BIOS setting.

Case is a Nexus Prominent 9. I really don't want to have to take the motherboard out to upgrade the HSF since it seems that the backplate needs to be changed on all the aftermarket ones. Weird...

 

I do hear the fan ramp up and get loud sometimes as I was transferring files to the backup server and while it was also unpacking some files

 

Anything specific to check in the BIOS @Squid? I am digging out my monitor and keyboard to check since it's running headless 

 

The hard drive are about 49-54 degrees C, when in use. Don't know if that helps?

Edited April 21, 20197 yr by maxse

30 minutes ago, maxse said:

The hard drive are about 49-54 degrees C

While within spec, IMO that's also hot.  I usually strive for max 39, and usually hit max 34 using 4 in 3 cases.

 

If you haven't added in an exhaust fan to that case (quick look shows that it has provisions, but doesn't come with one), you want to do that.

I looked at that case via Google.  What you want in any server is for all of the cooling air to come in the front of the case (across the disks) and exit out the rear of the case.  First thing, you want is a exhaust fan in the fan opening in the rear of that case.  Next thing you want to do is to block/close that fan opening in the bottom of the case. (You could put an exhaust fan in that opening but I don't think you could get much air flow by doing so because of the back pressure from surface that the case sets on.  Of course, you could cut a hole in that surface...  🙄 )

So I opened the case to check the fan situation.

 

Turns out the CPU fan was loose! One of the push-pin connectors somehow got loose and the HSF was basically dangling! I pushed it back into the hole and turned the pin again. Is this okay today? Or is the thermal past now messed up? I'm thinking I should just get a new HSF even though I'll have to take the motherboard out to change the plate.

 

Also, the case currently has an exhaust fan at the top, and an exhaust fan in the back. Which was working but now when I powered the server back up, I see it's not moving, so I'm thinking it's dead and will get a new fan for the back also...

 

The exhaust on the bottom of the case is basically blocked by my SSD that I just placed at the bottom. Any thoughts?

Do you guys know if I could use the Arctic Freezer 12 CO in this case with my micro ATX mobo?

 

*BTW I just checked the log after rebooting and saw this error. Is this anything relevant?

Apr 21 15:53:25 Sabertooth ntpd[1785]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
Apr 21 15:56:55 Sabertooth emhttpd: error: cmd: wrong csrf_token

Edited April 21, 20197 yr by maxse

17 minutes ago, maxse said:

*BTW I just checked the log after rebooting and saw this error. Is this anything relevant?

Apr 21 15:53:25 Sabertooth ntpd[1785]: kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
Apr 21 15:56:55 Sabertooth emhttpd: error: cmd: wrong csrf_token

Normal

17 minutes ago, maxse said:

Or is the thermal past now messed up?

Probably.  Paste is always a one-use thing and should be changed whenever a HSF is removed (which yours basically was).

20 minutes ago, maxse said:

One of the push-pin connectors somehow got loose

Yeah, Intel's direction arrows on the pins are counter-intuitive.  But, on the upside there have always been reports of computers such as Dell, etc where the HSF falls off during shipping due to the shitty design.