Rebooted, now power usage is +30W

[sremick]

Well this is odd. I've been running FreeNAS for almost 2 years now, through many reboots. The entire time I've been on 9.2.x (and still am). Had a power outage the other day (common) and my UPS did its traditional thing of signaling FreeNAS to shut down gracefully after 5 minutes. Everything worked fine and as designed. I then brought things back up, and everything worked fine still... but I noticed that the power meter on my UPS now showed the FreeNAS box using about 76W idle versus the 40-45W it had always used before. Very strange. I checked for runaway processes using top... I also tried another graceful reboot. Didn't help.

Any thoughts? What should I be looking at? Full hardware specs are in my signature.

 

[Mirfster]

Posted your specs since those on mobile devices don't see signatures:

FreeNAS: 9.2.1.9
Motherboard: ASRock E3C226D2I Mini-ITX Server Motherboard
CPU: Intel Xeon E3-1271 v3 Quad-Core 3.6GHz 8MB LGA 1150
CPU cooler: Thermalright TRUE Spirit 120M(BW) Rev.A
RAM: Crucial 16GB kit (8GBx2) DDR3 PC3-12800 Unbuffered ECC 1.35V CT2KIT102472BD160B
HDDs: 6x WD Red NAS 3TB IntelliPower 64MB Cache SATA 6.0Gb/s 3.5" WD30EFRX (RAIDZ2)
Power Supply: SeaSonic Platinum Series SS-400FL2 Active PFC F3 400W ATX12V Fanless 80 PLUS Platinum Certified
Case: Fractal Design Node 304

 

[m0nkey_]

You may need to adjust your lowest C state. I typically have to set mine to C3 and enable powerd. This usually saves me around 8 watts.

Code:

sysctl dev.cpu.0.cx_lowest=C3

You can set the above setting via a post-init command.

 

[sremick]

What would've changed during a reboot, when many other reboots (also on this same FreeNAS version with no changes/updates) had no effect to wattage usage?

 

[DrKK]

Are you sure the 30W is not your UPS recharging its battery? Sounds about right??

 

[sremick]

If that was the case it would've returned to normal within a few hours. Weeks later it's still drawing more watts idle post-reboot than it has in 2 years.

 

[Mirfster]

Got bad batteries in your UPS?

 

[sremick]

Got bad batteries in your UPS?

Doubt it. UPS is newer than the FreeNAS box (<2y) and display shows battery is healthy. Maybe if we were at 4+ years I'd expect the battery to be bad. But also that has nothing to do with how much power is flowing through the system, which is what the watt meter is measuring. While line power is present, the battery isn't in use. I could remove the battery and things would work.

 

[Nick2253]

Just to be clear, you're seeing this power draw according to your UPS's software? Do you have a way to independently verify this? Though unlikely, a powersurge could have damaged the controller in your UPS, which is causing it to report more power draw than otherwise. In my opinion, it seems much more likely that there is a problem with the UPS (either its hardware, software, or controller) than your FreeNAS system suddenly drawing more power.

 

[sremick]

Just to be clear, you're seeing this power draw according to your UPS's software? Do you have a way to independently verify this? Though unlikely, a powersurge could have damaged the controller in your UPS, which is causing it to report more power draw than otherwise. In my opinion, it seems much more likely that there is a problem with the UPS (either its hardware, software, or controller) than your FreeNAS system suddenly drawing more power.

Not UPS software: the UPS itself has a digital information display on the front, one of the bits of info is how many watts are being drawn by connected devices. And it wasn't a sudden surge... it was simply a reboot of the FreeNAS box.

I do have a separate Kill-A-Watt I can use to get a second opinion if you wish.

 

[Nick2253]

Not UPS software: the UPS itself has a digital information display on the front, one of the bits of info is how many watts are being drawn by connected devices. And it wasn't a sudden surge... it was simply a reboot of the FreeNAS box.

What kind of UPS is it? Also--and I know this is going to sound stupid--do you have anything else plugged in to the UPS?

Based on what you said earlier, my understanding is: you had a power outage, which caused the FreeNAS system to shut down cleanly, and only after you brought the FreeNAS system back up, did you suddenly notice that your power usage was not what it was before. In that case, a power outage could have easily been preceded by a power surge. Or, when the grid was brought back online, you experienced a power surge. It doesn't have to be huge to break electronics, or at least screw them up. Depending on how robust your PSU is, the simple act of turning on and off components connected to it could be enough to damage whatever sensor is being used to measure power draw.

I do have a separate Kill-A-Watt I can use to get a second opinion if you wish.

A second opinion would be valuable. You could even put the Kill-A-Watt inline with the UPS to compare power draw, and verify the numbers are accurate.

 

[DrKK]

The amount of watts difference we are talking about is considerable. Seems unlikely to be something in the OS drawing down more juice. Not with this board. You'd need a whole shitload of fans going at full blast to make 30 watts out of something. I'm just really at a loss. I want to say it has to be instrumentation error on whatever is measuring the watts?