I have a dual CPU Gold 6148 in an S2600STB motherboard, using an Intel case (Intel System Cabinet P4304XXMUXX Black) and Intel heat sinks,
Intel PSU (and 512GB RAM)
The system runs hot, around 60C when next to idle.
I replaced the paste between the CPU and the Heat Sink on one CPU and it made no difference.
I have noticed that during startup it spins up and then down the fans in the power supply, then spins up and then down the fans in the case.
However, I've never heard the case fans making a lot of noise while the system is running.
If I try to use a few CPUs at 100% it will heat up to 95C and any more then it will shut down.
However, the fans wont spin up faster when this happens.
I'm using latest Ubuntu 16.04.4 LTS
It would seem that this doesn't support the motherboard by default since the fans wont spin up to cool it down.
How do I resolve this?
I am sorry for missing Ubuntu support. Also with the system fan not working as they should.
Let's start by gathering some general information.
Please flash BIOS to latest - https://downloadcenter.intel.com/product/124477/Intel-Server-Board-S2600STB.
Also, you have mentioned that you also have the P4304XXMUXX chassis. Also, what configuration of Fans are you using.
If there is anything else we can help please feel free to ask.
This server is about 7 months old (and yes been pretty idle for 7 months)
The BIOS was the oldest version 1.2001 so I had to go to a sizeable amount of trouble to upgrade it to the latest BIOS (three upgrades: 4 then 9 then 13)
Alas this has now made it worse.
The CPU temperatures are higher after boot: 75C on CPU0 and 95C on CPU1
instead of the usual 60C on both
The fans are the default 5 fans that came in the Intel case, that are, of course, powered from the motherboard, and in between the HDD enclosures at the front and the motherboard at the back.
... any other suggestions?
Hi, yep tried "Performance" now also.
The CPUs still heated up and had alarms/critical alerts on the console, linux said it slowed the cores down but eventually it shut down after about half an hour.
I pushed the CPUs - using about 25 threads (of the 80) and CPU1 first went all cores to 95C then CPU0 went all cores to 90C-94C before it shut down.
The fans didn't make a sound.
As I mentioned before, during a full power cycle I do hear the PSU fan and then the case fans spinning up high then back down.
I've also checked that the case fans are actually the second sound by powering it up with the side open so I can see them spinning and that they're the source of the 2nd louder fan noise during power up.
I next rebooted it again after letting it cool down, this time I set "Fan PWM Offset" to 50 thus the case fans should always be at least at 50% ?
But, still no obvious sounds from the case fans - even when pushing the temperature up to 83C/94C by using a few cores - which I stopped since I don't like having it power off when running.
Since the fans do actually spin up when the server first powers on, I presume they should actually work (loudly) when the temp gets above high (85C) or close to critical (95C)
Any other suggestions?
According to the logs we have some issues that can be causing this problem.
The first thing that we need to check is the SDR version package that according to the log is not up to date and this could be causing the issue.
Then we need to delete the SEL file in the BIOS or setting the "Processor Retest" in order to delete the SEL file.
Please check that the processor has enough thermal grease between the processor and the heatsinks and check that the heatsinks completely screwed down on to the system board.
Please perform this actions and let us know how it goes.
the problem was indeed the SDR in the BIOS.
I'm not sure why it was D1.43 - or even where version D1.43 came from since it's not listed in any of the BIOS updates, nor why the last Startup.nsh skipped the SDR update, or why the original motherboard BIOS didn't work with the fans at all,
but directly installing SDR 1.28 from the boot UEFI shell, using the updS2600STB.nsh in 1.0013 made the fans immediately come to life at the 50% PWM setting I had in the BIOS.
Thanks for finding the BIOS problem.